diff --git a/.github/dependabot.yml b/.github/dependabot.yml
index c43a610fa2c24..da99e7a0a0d23 100644
--- a/.github/dependabot.yml
+++ b/.github/dependabot.yml
@@ -22,6 +22,9 @@ updates:
       update-types: ["version-update:semver-minor", "version-update:semver-major"]
     - dependency-name: "parquet"
       update-types: ["version-update:semver-minor", "version-update:semver-major"]
+    # bump sqllogictest manually together with sqllogictest-bin in CI docker image
+    - dependency-name: "sqllogictest"
+      update-types: ["version-update:semver-minor", "version-update:semver-major"]
   # Create a group of dependencies to be updated together in one pull request
   groups:
     aws:
diff --git a/Cargo.lock b/Cargo.lock
index b5f04b2a910bb..fee7254124c22 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1331,22 +1331,6 @@ dependencies = [
  "zeroize",
 ]
 
-[[package]]
-name = "aws-http"
-version = "0.60.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "30e4199d5d62ab09be6a64650c06cc5c4aa45806fed4c74bc4a5c8eaf039a6fa"
-dependencies = [
- "aws-smithy-runtime-api",
- "aws-smithy-types",
- "aws-types",
- "bytes",
- "http 0.2.9",
- "http-body 0.4.5",
- "pin-project-lite",
- "tracing",
-]
-
 [[package]]
 name = "aws-lc-rs"
 version = "1.6.2"
@@ -1597,7 +1581,7 @@ dependencies = [
  "hex",
  "hmac",
  "http 0.2.9",
- "http 1.1.0",
+ "http 1.2.0",
  "once_cell",
  "p256 0.11.1",
  "percent-encoding",
@@ -1729,7 +1713,7 @@ dependencies = [
  "aws-smithy-types",
  "bytes",
  "http 0.2.9",
- "http 1.1.0",
+ "http 1.2.0",
  "pin-project-lite",
  "tokio",
  "tracing",
@@ -1747,7 +1731,7 @@ dependencies = [
  "bytes-utils",
  "futures-core",
  "http 0.2.9",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 0.4.5",
  "http-body 1.0.0",
  "http-body-util",
@@ -1806,7 +1790,7 @@ dependencies = [
  "axum-core",
  "bytes",
  "futures-util",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "http-body-util",
  "hyper 1.4.1",
@@ -1839,7 +1823,7 @@ dependencies = [
  "async-trait",
  "bytes",
  "futures-util",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "http-body-util",
  "mime",
@@ -1859,7 +1843,7 @@ checksum = "077959a7f8cf438676af90b483304528eb7e16eadadb7f44e9ada4f9dceb9e62"
 dependencies = [
  "axum-core",
  "chrono",
- "http 1.1.0",
+ "http 1.2.0",
  "mime_guess",
  "rust-embed",
  "tower-service",
@@ -1875,7 +1859,7 @@ dependencies = [
  "axum-core",
  "bytes",
  "futures-util",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "http-body-util",
  "mime",
@@ -3506,19 +3490,6 @@ version = "0.3.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7762d17f1241643615821a8455a0b2c3e803784b058693d990b11f2dce25a0ca"
 
-[[package]]
-name = "dashmap"
-version = "5.5.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "978747c1d849a7d2ee5e8adc0159961c48fb7e5db2f06af6723b80123bb53856"
-dependencies = [
- "cfg-if",
- "hashbrown 0.14.5",
- "lock_api",
- "once_cell",
- "parking_lot_core 0.9.8",
-]
-
 [[package]]
 name = "dashmap"
 version = "6.1.0"
@@ -3561,7 +3532,7 @@ dependencies = [
  "bytes",
  "bzip2",
  "chrono",
- "dashmap 6.1.0",
+ "dashmap",
  "datafusion-catalog",
  "datafusion-common",
  "datafusion-common-runtime",
@@ -3654,7 +3625,7 @@ checksum = "799e70968c815b611116951e3dd876aef04bf217da31b72eec01ee6a959336a1"
 dependencies = [
  "arrow 52.2.0",
  "chrono",
- "dashmap 6.1.0",
+ "dashmap",
  "datafusion-common",
  "datafusion-expr",
  "futures",
@@ -4050,7 +4021,7 @@ dependencies = [
  "bytes",
  "cfg-if",
  "chrono",
- "dashmap 6.1.0",
+ "dashmap",
  "datafusion",
  "datafusion-common",
  "datafusion-expr",
@@ -5346,21 +5317,6 @@ dependencies = [
  "libc",
 ]
 
-[[package]]
-name = "function_name"
-version = "0.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b1ab577a896d09940b5fe12ec5ae71f9d8211fff62c919c03a3750a9901e98a7"
-dependencies = [
- "function_name-proc-macro",
-]
-
-[[package]]
-name = "function_name-proc-macro"
-version = "0.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "673464e1e314dd67a0fd9544abc99e8eb28d0c7e3b69b033bcff9b2d00b87333"
-
 [[package]]
 name = "funty"
 version = "2.0.0"
@@ -5646,7 +5602,7 @@ checksum = "cae77099e2399aea466bba05f0d23a150b4f34ed7ce535835e71d91399e65b58"
 dependencies = [
  "anyhow",
  "async-trait",
- "http 1.1.0",
+ "http 1.2.0",
  "thiserror 1.0.63",
  "tokio",
  "tonic",
@@ -5729,7 +5685,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9c3eaaad103912825594d674a4b1e556ccbb05a13a6cac17dcfd871997fb760a"
 dependencies = [
  "google-cloud-token",
- "http 1.1.0",
+ "http 1.2.0",
  "thiserror 1.0.63",
  "tokio",
  "tokio-retry",
@@ -5795,14 +5751,12 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "68a7f542ee6b35af73b06abc0dad1c1bae89964e4e253bc4b587b91c9637867b"
 dependencies = [
  "cfg-if",
- "dashmap 5.5.3",
  "futures",
  "futures-timer",
  "no-std-compat",
  "nonzero_ext",
  "parking_lot 0.12.1",
  "portable-atomic",
- "rand",
  "smallvec",
  "spinning_top",
 ]
@@ -5859,7 +5813,7 @@ dependencies = [
  "fnv",
  "futures-core",
  "futures-sink",
- "http 1.1.0",
+ "http 1.2.0",
  "indexmap 2.7.0",
  "slab",
  "tokio",
@@ -6058,9 +6012,9 @@ dependencies = [
 
 [[package]]
 name = "http"
-version = "1.1.0"
+version = "1.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "21b9ddb458710bc376481b842f5da65cdf31522de232c1ca8146abce2a358258"
+checksum = "f16ca2af56261c99fba8bac40a10251ce8188205a4c448fbb745a2e4daa76fea"
 dependencies = [
  "bytes",
  "fnv",
@@ -6085,7 +6039,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1cac85db508abc24a2e48553ba12a996e87244a0395ce011e62b37158745d643"
 dependencies = [
  "bytes",
- "http 1.1.0",
+ "http 1.2.0",
 ]
 
 [[package]]
@@ -6096,7 +6050,7 @@ checksum = "0475f8b2ac86659c21b64320d5d653f9efe42acd2a4e560073ec61a155a34f1d"
 dependencies = [
  "bytes",
  "futures-core",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "pin-project-lite",
 ]
@@ -6159,7 +6113,7 @@ dependencies = [
  "futures-channel",
  "futures-util",
  "h2 0.4.7",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "httparse",
  "httpdate",
@@ -6194,7 +6148,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a0bea761b46ae2b24eb4aef630d8d1c398157b6fc29e6350ecf090a0b70c952c"
 dependencies = [
  "futures-util",
- "http 1.1.0",
+ "http 1.2.0",
  "hyper 1.4.1",
  "hyper-util",
  "rustls 0.22.4",
@@ -6255,7 +6209,7 @@ dependencies = [
  "bytes",
  "futures-channel",
  "futures-util",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "hyper 1.4.1",
  "pin-project-lite",
@@ -6373,7 +6327,7 @@ source = "git+https://github.com/risingwavelabs/iceberg-rust.git?rev=53f786fb214
 dependencies = [
  "async-trait",
  "chrono",
- "http 1.1.0",
+ "http 1.2.0",
  "iceberg",
  "itertools 0.13.0",
  "log",
@@ -7365,7 +7319,7 @@ dependencies = [
  "aws-smithy-types",
  "aws-types",
  "bytes",
- "http 1.1.0",
+ "http 1.2.0",
  "madsim",
  "spin 0.9.8",
  "tracing",
@@ -7442,7 +7396,7 @@ checksum = "f271a476bbaa9d2139e1e1a5beb869c6119e805a0b67ad2b2857e4a8785b111a"
 dependencies = [
  "prettyplease 0.2.15",
  "proc-macro2",
- "prost-build 0.13.1",
+ "prost-build 0.13.4",
  "quote",
  "syn 2.0.87",
  "tonic-build",
@@ -8312,7 +8266,7 @@ dependencies = [
  "flagset",
  "futures",
  "getrandom",
- "http 1.1.0",
+ "http 1.2.0",
  "log",
  "md-5",
  "once_cell",
@@ -8344,7 +8298,7 @@ dependencies = [
  "flagset",
  "futures",
  "getrandom",
- "http 1.1.0",
+ "http 1.2.0",
  "log",
  "md-5",
  "once_cell",
@@ -8486,7 +8440,7 @@ checksum = "6b925a602ffb916fb7421276b86756027b37ee708f9dce2dbdcc51739f07e727"
 dependencies = [
  "async-trait",
  "futures-core",
- "http 1.1.0",
+ "http 1.2.0",
  "opentelemetry",
  "opentelemetry-proto",
  "opentelemetry_sdk",
@@ -9594,6 +9548,15 @@ dependencies = [
  "prost-derive 0.13.1",
 ]
 
+[[package]]
+name = "prost"
+version = "0.13.4"
+source = "git+https://github.com/xxchan/prost.git?rev=0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163#0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163"
+dependencies = [
+ "bytes",
+ "prost-derive 0.13.4",
+]
+
 [[package]]
 name = "prost-build"
 version = "0.11.9"
@@ -9618,20 +9581,18 @@ dependencies = [
 
 [[package]]
 name = "prost-build"
-version = "0.13.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5bb182580f71dd070f88d01ce3de9f4da5021db7115d2e1c3605a754153b77c1"
+version = "0.13.4"
+source = "git+https://github.com/xxchan/prost.git?rev=0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163#0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163"
 dependencies = [
- "bytes",
- "heck 0.5.0",
- "itertools 0.13.0",
+ "heck 0.4.1",
+ "itertools 0.10.5",
  "log",
- "multimap 0.10.0",
+ "multimap 0.8.3",
  "once_cell",
  "petgraph",
  "prettyplease 0.2.15",
- "prost 0.13.1",
- "prost-types 0.13.1",
+ "prost 0.13.4",
+ "prost-types 0.13.4",
  "regex",
  "syn 2.0.87",
  "tempfile",
@@ -9670,7 +9631,19 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "18bec9b0adc4eba778b33684b7ba3e7137789434769ee3ce3930463ef904cfca"
 dependencies = [
  "anyhow",
- "itertools 0.13.0",
+ "itertools 0.10.5",
+ "proc-macro2",
+ "quote",
+ "syn 2.0.87",
+]
+
+[[package]]
+name = "prost-derive"
+version = "0.13.4"
+source = "git+https://github.com/xxchan/prost.git?rev=0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163#0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163"
+dependencies = [
+ "anyhow",
+ "itertools 0.10.5",
  "proc-macro2",
  "quote",
  "syn 2.0.87",
@@ -9719,6 +9692,14 @@ dependencies = [
  "prost 0.13.1",
 ]
 
+[[package]]
+name = "prost-types"
+version = "0.13.4"
+source = "git+https://github.com/xxchan/prost.git?rev=0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163#0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163"
+dependencies = [
+ "prost 0.13.4",
+]
+
 [[package]]
 name = "protobuf"
 version = "2.28.0"
@@ -10234,7 +10215,7 @@ dependencies = [
  "hex",
  "hmac",
  "home",
- "http 1.1.0",
+ "http 1.2.0",
  "jsonwebtoken",
  "log",
  "once_cell",
@@ -10306,7 +10287,7 @@ dependencies = [
  "futures-core",
  "futures-util",
  "h2 0.4.7",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "http-body-util",
  "hyper 1.4.1",
@@ -10353,7 +10334,7 @@ checksum = "a45d100244a467870f6cb763c4484d010a6bed6bd610b3676e3825d93fb4cfbd"
 dependencies = [
  "anyhow",
  "async-trait",
- "http 1.1.0",
+ "http 1.2.0",
  "reqwest 0.12.4",
  "serde",
  "thiserror 1.0.63",
@@ -10455,7 +10436,6 @@ dependencies = [
  "regex",
  "reqwest 0.12.4",
  "serde",
- "serde_json",
  "serde_with 3.8.1",
  "serde_yaml",
  "sqlx",
@@ -10499,7 +10479,6 @@ version = "2.3.0-alpha"
 dependencies = [
  "anyhow",
  "async-trait",
- "bincode 1.3.3",
  "bytes",
  "itertools 0.13.0",
  "parking_lot 0.12.1",
@@ -10522,7 +10501,6 @@ dependencies = [
  "anyhow",
  "async-recursion",
  "async-trait",
- "criterion",
  "either",
  "futures",
  "futures-async-stream",
@@ -10552,7 +10530,6 @@ dependencies = [
  "tempfile",
  "thiserror 1.0.63",
  "thiserror-ext",
- "tikv-jemallocator",
  "tokio-postgres",
  "tokio-stream 0.1.15",
  "tracing",
@@ -10654,7 +10631,6 @@ version = "2.3.0-alpha"
 dependencies = [
  "clap",
  "madsim-tokio",
- "prometheus",
  "risingwave_batch_executors",
  "risingwave_common",
  "risingwave_compactor",
@@ -10673,14 +10649,12 @@ dependencies = [
 name = "risingwave_cmd_all"
 version = "2.3.0-alpha"
 dependencies = [
- "anyhow",
  "clap",
  "console",
  "const-str",
  "expect-test",
  "home",
  "madsim-tokio",
- "prometheus",
  "risingwave_batch_executors",
  "risingwave_cmd",
  "risingwave_common",
@@ -10745,10 +10719,9 @@ dependencies = [
  "governor",
  "hashbrown 0.14.5",
  "hex",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 0.4.5",
  "humantime",
- "hytra",
  "itertools 0.13.0",
  "itoa",
  "jsonbb",
@@ -10829,7 +10802,6 @@ dependencies = [
  "ethnum",
  "fixedbitset 0.5.0",
  "jsonbb",
- "lru 0.7.6",
  "risingwave_common_proc_macro",
  "rust_decimal",
  "serde_json",
@@ -10839,7 +10811,6 @@ dependencies = [
 name = "risingwave_common_heap_profiling"
 version = "2.3.0-alpha"
 dependencies = [
- "anyhow",
  "chrono",
  "madsim-tokio",
  "parking_lot 0.12.1",
@@ -10862,7 +10833,7 @@ dependencies = [
  "easy-ext",
  "futures",
  "http 0.2.9",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "hyper 0.14.27",
  "hyper 1.4.1",
@@ -10924,7 +10895,7 @@ dependencies = [
  "axum",
  "axum-extra",
  "futures",
- "http 1.1.0",
+ "http 1.2.0",
  "madsim-tokio",
  "madsim-tonic",
  "prometheus",
@@ -10945,21 +10916,14 @@ name = "risingwave_compaction_test"
 version = "2.3.0-alpha"
 dependencies = [
  "anyhow",
- "async-trait",
  "bytes",
  "clap",
  "foyer",
- "futures",
  "madsim-tokio",
- "prometheus",
- "rand",
  "risingwave_common",
  "risingwave_compactor",
  "risingwave_hummock_sdk",
- "risingwave_hummock_test",
- "risingwave_meta",
  "risingwave_meta_node",
- "risingwave_object_store",
  "risingwave_pb",
  "risingwave_rpc_client",
  "risingwave_rt",
@@ -10978,7 +10942,6 @@ dependencies = [
  "jsonbb",
  "madsim-tokio",
  "madsim-tonic",
- "parking_lot 0.12.1",
  "prost 0.13.1",
  "risingwave_common",
  "risingwave_common_heap_profiling",
@@ -11005,8 +10968,7 @@ dependencies = [
  "foyer",
  "futures",
  "futures-async-stream",
- "http 1.1.0",
- "hyper 1.4.1",
+ "http 1.2.0",
  "itertools 0.13.0",
  "madsim-tokio",
  "madsim-tonic",
@@ -11036,7 +10998,6 @@ dependencies = [
  "tokio-stream 0.1.15",
  "tower 0.5.0",
  "tracing",
- "uuid",
  "workspace-hack",
 ]
 
@@ -11090,7 +11051,6 @@ dependencies = [
  "google-cloud-googleapis",
  "google-cloud-pubsub",
  "governor",
- "http 0.2.9",
  "iceberg",
  "iceberg-catalog-glue",
  "iceberg-catalog-rest",
@@ -11098,8 +11058,6 @@ dependencies = [
  "indexmap 2.7.0",
  "itertools 0.13.0",
  "jni",
- "jsonbb",
- "jsonwebtoken",
  "madsim-rdkafka",
  "madsim-tokio",
  "madsim-tonic",
@@ -11134,7 +11092,6 @@ dependencies = [
  "risingwave_common_estimate_size",
  "risingwave_connector_codec",
  "risingwave_jni_core",
- "risingwave_object_store",
  "risingwave_pb",
  "risingwave_rpc_client",
  "rumqttc",
@@ -11194,7 +11151,7 @@ dependencies = [
  "madsim-tokio",
  "num-bigint",
  "prost 0.13.1",
- "prost-build 0.13.1",
+ "prost-build 0.13.4",
  "prost-reflect",
  "prost-types 0.13.1",
  "protox",
@@ -11225,7 +11182,6 @@ dependencies = [
  "itertools 0.13.0",
  "madsim-tokio",
  "madsim-tonic",
- "memcomparable",
  "prost 0.13.1",
  "regex",
  "risingwave_common",
@@ -11234,7 +11190,6 @@ dependencies = [
  "risingwave_hummock_sdk",
  "risingwave_meta",
  "risingwave_meta_model",
- "risingwave_meta_model_migration",
  "risingwave_object_store",
  "risingwave_pb",
  "risingwave_rpc_client",
@@ -11256,21 +11211,15 @@ name = "risingwave_dml"
 version = "2.3.0-alpha"
 dependencies = [
  "assert_matches",
- "criterion",
  "futures",
  "futures-async-stream",
  "itertools 0.13.0",
  "madsim-tokio",
  "parking_lot 0.12.1",
  "paste",
- "rand",
  "risingwave_common",
- "risingwave_connector",
- "risingwave_pb",
- "rw_futures_util",
  "tempfile",
  "thiserror 1.0.63",
- "thiserror-ext",
  "tracing",
  "workspace-hack",
 ]
@@ -11296,7 +11245,6 @@ version = "2.3.0-alpha"
 dependencies = [
  "anyhow",
  "bincode 1.3.3",
- "bytes",
  "easy-ext",
  "madsim-tonic",
  "serde",
@@ -11314,7 +11262,6 @@ dependencies = [
  "async-trait",
  "auto_impl",
  "await-tree",
- "cfg-or-panic",
  "chrono",
  "const-currying",
  "downcast-rs",
@@ -11476,7 +11423,6 @@ dependencies = [
  "risingwave_expr_impl",
  "risingwave_frontend_macro",
  "risingwave_hummock_sdk",
- "risingwave_object_store",
  "risingwave_pb",
  "risingwave_rpc_client",
  "risingwave_sqlparser",
@@ -11500,7 +11446,6 @@ dependencies = [
  "url",
  "uuid",
  "workspace-hack",
- "zstd 0.13.0",
 ]
 
 [[package]]
@@ -11518,16 +11463,13 @@ version = "2.3.0-alpha"
 dependencies = [
  "bincode 2.0.0-rc.3",
  "bytes",
- "easy-ext",
  "hex",
  "itertools 0.13.0",
  "parse-display",
- "prost 0.13.1",
  "risingwave_common",
  "risingwave_common_estimate_size",
  "risingwave_pb",
  "serde",
- "serde_bytes",
  "tracing",
  "workspace-hack",
 ]
@@ -11559,7 +11501,6 @@ dependencies = [
  "risingwave_rpc_client",
  "risingwave_storage",
  "risingwave_test_runner",
- "serde",
  "serial_test",
  "sync-point",
  "workspace-hack",
@@ -11622,21 +11563,14 @@ dependencies = [
  "cfg-or-panic",
  "chrono",
  "expect-test",
- "foyer",
  "fs-err",
  "futures",
- "itertools 0.13.0",
  "jni",
  "madsim-tokio",
  "paste",
  "prost 0.13.1",
  "risingwave_common",
- "risingwave_expr",
- "risingwave_hummock_sdk",
  "risingwave_pb",
- "rw_futures_util",
- "serde",
- "serde_json",
  "thiserror 1.0.63",
  "thiserror-ext",
  "tracing",
@@ -11661,15 +11595,10 @@ dependencies = [
 name = "risingwave_mem_table_spill_test"
 version = "2.3.0-alpha"
 dependencies = [
- "async-trait",
- "bytes",
- "futures",
- "futures-async-stream",
  "madsim-tokio",
  "risingwave_common",
  "risingwave_hummock_sdk",
  "risingwave_hummock_test",
- "risingwave_storage",
  "risingwave_stream",
 ]
 
@@ -11681,10 +11610,8 @@ dependencies = [
  "arc-swap",
  "assert_matches",
  "async-trait",
- "aws-config",
  "axum",
  "base64-url",
- "bincode 1.3.3",
  "bytes",
  "chrono",
  "clap",
@@ -11696,19 +11623,14 @@ dependencies = [
  "enum-as-inner 0.6.0",
  "expect-test",
  "fail",
- "flate2",
- "function_name",
  "futures",
  "hex",
- "http 1.1.0",
+ "http 1.2.0",
  "itertools 0.13.0",
  "jsonbb",
  "madsim-tokio",
  "madsim-tonic",
  "maplit",
- "memcomparable",
- "mime_guess",
- "moka",
  "notify",
  "num-integer",
  "num-traits",
@@ -11732,7 +11654,6 @@ dependencies = [
  "risingwave_rpc_client",
  "risingwave_sqlparser",
  "risingwave_test_runner",
- "rw-aws-sdk-ec2",
  "rw_futures_util",
  "scopeguard",
  "sea-orm",
@@ -11749,7 +11670,6 @@ dependencies = [
  "tower-http",
  "tracing",
  "tracing-subscriber",
- "url",
  "uuid",
  "workspace-hack",
 ]
@@ -11781,7 +11701,6 @@ version = "2.3.0-alpha"
 dependencies = [
  "prost 0.13.1",
  "risingwave_common",
- "risingwave_hummock_sdk",
  "risingwave_pb",
  "sea-orm",
  "serde",
@@ -11805,13 +11724,9 @@ dependencies = [
 name = "risingwave_meta_node"
 version = "2.3.0-alpha"
 dependencies = [
- "anyhow",
  "clap",
  "educe",
- "either",
- "futures",
  "hex",
- "itertools 0.13.0",
  "madsim-tokio",
  "madsim-tonic",
  "otlp-embedded",
@@ -11822,12 +11737,10 @@ dependencies = [
  "risingwave_common_heap_profiling",
  "risingwave_common_service",
  "risingwave_meta",
- "risingwave_meta_model_migration",
  "risingwave_meta_service",
  "risingwave_pb",
  "risingwave_rpc_client",
  "sea-orm",
- "serde",
  "serde_json",
  "thiserror-ext",
  "tracing",
@@ -11840,12 +11753,10 @@ version = "2.3.0-alpha"
 dependencies = [
  "anyhow",
  "async-trait",
- "either",
  "futures",
  "itertools 0.13.0",
  "madsim-tokio",
  "madsim-tonic",
- "prost 0.13.1",
  "rand",
  "regex",
  "risingwave_common",
@@ -11856,7 +11767,6 @@ dependencies = [
  "risingwave_pb",
  "sea-orm",
  "serde_json",
- "sync-point",
  "thiserror-ext",
  "tokio-stream 0.1.15",
  "tracing",
@@ -11877,12 +11787,10 @@ dependencies = [
  "aws-smithy-types",
  "bytes",
  "crc32fast",
- "either",
  "fail",
  "futures",
  "hyper 0.14.27",
  "hyper-rustls 0.24.2",
- "hyper-tls 0.5.0",
  "itertools 0.13.0",
  "madsim",
  "madsim-aws-sdk-s3",
@@ -11910,10 +11818,11 @@ dependencies = [
  "pbjson",
  "pbjson-build",
  "prost 0.13.1",
- "prost-build 0.13.1",
+ "prost-build 0.13.4",
  "prost-helpers",
  "risingwave_error",
  "serde",
+ "static_assertions",
  "strum 0.26.3",
  "thiserror 1.0.63",
  "walkdir",
@@ -11965,10 +11874,7 @@ dependencies = [
  "easy-ext",
  "either",
  "futures",
- "h2 0.4.7",
- "http 1.1.0",
- "hyper 1.4.1",
- "itertools 0.13.0",
+ "http 1.2.0",
  "lru 0.7.6",
  "madsim-tokio",
  "madsim-tonic",
@@ -11987,7 +11893,6 @@ dependencies = [
  "tokio-stream 0.1.15",
  "tower 0.5.0",
  "tracing",
- "url",
  "workspace-hack",
 ]
 
@@ -12039,14 +11944,12 @@ dependencies = [
  "itertools 0.13.0",
  "lru 0.7.6",
  "madsim",
- "madsim-aws-sdk-s3",
  "madsim-rdkafka",
  "madsim-tokio",
  "maplit",
  "paste",
  "pin-project",
  "pretty_assertions",
- "prometheus",
  "rand",
  "rand_chacha",
  "risingwave_batch_executors",
@@ -12062,7 +11965,6 @@ dependencies = [
  "risingwave_meta_node",
  "risingwave_object_store",
  "risingwave_pb",
- "risingwave_rpc_client",
  "risingwave_sqlparser",
  "risingwave_sqlsmith",
  "serde",
@@ -12072,7 +11974,6 @@ dependencies = [
  "tempfile",
  "tikv-jemallocator",
  "tokio-postgres",
- "tokio-stream 0.1.15",
  "tracing",
  "tracing-subscriber",
  "uuid",
@@ -12117,7 +12018,6 @@ dependencies = [
  "risingwave_expr",
  "risingwave_expr_impl",
  "risingwave_frontend",
- "risingwave_pb",
  "risingwave_sqlparser",
  "similar",
  "thiserror-ext",
@@ -12135,7 +12035,6 @@ dependencies = [
  "clap",
  "futures",
  "madsim-tokio",
- "prometheus",
  "regex",
  "risingwave_rt",
  "serde",
@@ -12151,7 +12050,6 @@ dependencies = [
 name = "risingwave_storage"
 version = "2.3.0-alpha"
 dependencies = [
- "ahash 0.8.11",
  "anyhow",
  "arc-swap",
  "async-trait",
@@ -12160,9 +12058,7 @@ dependencies = [
  "bincode 1.3.3",
  "bytes",
  "criterion",
- "crossbeam",
  "darwin-libproc",
- "dashmap 6.1.0",
  "dyn-clone",
  "either",
  "enum-as-inner 0.6.0",
@@ -12277,7 +12173,6 @@ dependencies = [
  "serde_yaml",
  "smallvec",
  "static_assertions",
- "strum 0.26.3",
  "strum_macros 0.26.4",
  "thiserror 1.0.63",
  "thiserror-ext",
@@ -12299,10 +12194,8 @@ dependencies = [
  "prost 0.13.1",
  "reqwest 0.12.4",
  "risingwave_pb",
- "thiserror 1.0.63",
  "thiserror-ext",
  "tracing",
- "uuid",
 ]
 
 [[package]]
@@ -12726,30 +12619,6 @@ dependencies = [
  "wait-timeout",
 ]
 
-[[package]]
-name = "rw-aws-sdk-ec2"
-version = "1.2.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "80dba3602b267a7f9dcc546ccbf1d05752447773146253c7e344e2a320630b7f"
-dependencies = [
- "aws-credential-types",
- "aws-http",
- "aws-runtime",
- "aws-smithy-async",
- "aws-smithy-http",
- "aws-smithy-json",
- "aws-smithy-query",
- "aws-smithy-runtime",
- "aws-smithy-runtime-api",
- "aws-smithy-types",
- "aws-smithy-xml",
- "aws-types",
- "fastrand",
- "http 0.2.9",
- "regex",
- "tracing",
-]
-
 [[package]]
 name = "rw_futures_util"
 version = "0.0.0"
@@ -13775,9 +13644,9 @@ dependencies = [
 
 [[package]]
 name = "sqllogictest"
-version = "0.23.0"
+version = "0.24.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ec31dce96f489e2247a165837f49bbce4912b0cbcf127b79b4fdd87503022ae9"
+checksum = "48c03edcabfda1ab894cc63a115b9f014bfc6875916b850ab7498d3cb92daed9"
 dependencies = [
  "async-trait",
  "educe",
@@ -14768,7 +14637,7 @@ dependencies = [
  "bytes",
  "futures-core",
  "futures-sink",
- "http 1.1.0",
+ "http 1.2.0",
  "httparse",
  "rand",
  "ring 0.17.5",
@@ -14862,7 +14731,7 @@ dependencies = [
  "bytes",
  "flate2",
  "h2 0.4.7",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "http-body-util",
  "hyper 1.4.1",
@@ -14891,7 +14760,7 @@ checksum = "fe4ee8877250136bd7e3d2331632810a4df4ea5e004656990d8d66d2f5ee8a67"
 dependencies = [
  "prettyplease 0.2.15",
  "proc-macro2",
- "prost-build 0.13.1",
+ "prost-build 0.13.4",
  "quote",
  "syn 2.0.87",
 ]
@@ -14941,7 +14810,7 @@ dependencies = [
  "bytes",
  "futures-core",
  "futures-util",
- "http 1.1.0",
+ "http 1.2.0",
  "http-body 1.0.0",
  "http-body-util",
  "http-range-header",
@@ -16564,7 +16433,6 @@ dependencies = [
 name = "with_options"
 version = "2.3.0-alpha"
 dependencies = [
- "proc-macro2",
  "quote",
  "syn 2.0.87",
 ]
diff --git a/Cargo.toml b/Cargo.toml
index 4a86290bbc128..c7177b037f736 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -80,6 +80,19 @@ keywords = ["sql", "database", "streaming"]
 license = "Apache-2.0"
 repository = "https://github.com/risingwavelabs/risingwave"
 
+# some dependencies are not worth deleting.
+[workspace.metadata.cargo-machete]
+ignored = [
+    "workspace-hack",
+    "expect-test",
+    "pretty_assertions",
+    "serde",
+    "serde_json",
+]
+[workspace.metadata.cargo-udeps.ignore]
+normal = ["workspace-hack"]
+development = ["expect-test", "pretty_assertions"]
+
 [workspace.dependencies]
 foyer = { version = "0.13.1", features = ["tracing", "nightly", "prometheus"] }
 apache-avro = { git = "https://github.com/risingwavelabs/avro", rev = "25113ba88234a9ae23296e981d8302c290fdaa4b", features = [
@@ -107,11 +120,6 @@ aws-sdk-s3 = { version = "1", default-features = false, features = [
     "rt-tokio",
     "rustls",
 ] }
-# To bump the version of aws-sdk-ec2, check the README of https://github.com/risingwavelabs/rw-aws-sdk-ec2
-aws-sdk-ec2 = { package = "rw-aws-sdk-ec2", version = "1", default-features = false, features = [
-    "rt-tokio",
-    "rustls",
-] }
 aws-sdk-sqs = { version = "1", default-features = false, features = [
     "rt-tokio",
     "rustls",
@@ -204,6 +212,7 @@ tokio-stream = { git = "https://github.com/madsim-rs/tokio.git", rev = "0dd1055"
 tokio-util = "0.7"
 tracing-opentelemetry = "0.25"
 rand = { version = "0.8", features = ["small_rng"] }
+governor = { version = "0.6", default-features = false, features = ["std"] }
 risingwave_backup = { path = "./src/storage/backup" }
 risingwave_batch = { path = "./src/batch" }
 risingwave_batch_executors = { path = "./src/batch/executors" }
@@ -366,6 +375,8 @@ sqlx = { git = "https://github.com/madsim-rs/sqlx.git", rev = "3efe6d0065963db2a
 futures-timer = { git = "https://github.com/madsim-rs/futures-timer.git", rev = "05b33b4" }
 # patch to remove preserve_order from serde_json
 bson = { git = "https://github.com/risingwavelabs/bson-rust", rev = "e5175ec" }
+# TODO: unpatch after PR merged https://github.com/tokio-rs/prost/pull/1210
+prost-build = { git = "https://github.com/xxchan/prost.git", rev = "0eb1c7b09976cf6b5231e4b8d87bb5908ae6a163" }
 
 [workspace.metadata.dylint]
 libraries = [{ path = "./lints" }]
diff --git a/Makefile.toml b/Makefile.toml
index 89814dbe36ae5..6079a69e6ab4d 100644
--- a/Makefile.toml
+++ b/Makefile.toml
@@ -382,11 +382,7 @@ ln -s "$(pwd)/target/${RISEDEV_BUILD_TARGET_DIR}${BUILD_MODE_DIR}/risingwave" "$
 [tasks.post-build-risingwave]
 category = "RiseDev - Build"
 description = "Copy RisingWave binaries to bin"
-condition = { env_set = [
-    "ENABLE_BUILD_RUST",
-], env_not_set = [
-    "USE_SYSTEM_RISINGWAVE",
-] }
+condition = { env_set = ["ENABLE_BUILD_RUST"], env_not_set = ["USE_SYSTEM_RISINGWAVE"] }
 dependencies = [
     "link-all-in-one-binaries",
     "link-user-bin",
@@ -649,13 +645,23 @@ echo "   $(tput setaf 4)source ${PREFIX_CONFIG}/psql-env$(tput sgr0)"
 
 [tasks.psql]
 category = "RiseDev - Start/Stop"
-description = "Run local psql client with default connection parameters. You can pass extra arguments to psql."
+description = "Run local psql client for RisingWave with default connection parameters. You can pass extra arguments to psql."
 dependencies = ["check-and-load-risedev-env-file"]
 script = '''
 #!/usr/bin/env bash
 psql -h $RISEDEV_RW_FRONTEND_LISTEN_ADDRESS -p $RISEDEV_RW_FRONTEND_PORT -U root -d dev "$@"
 '''
 
+[tasks.pgpsql]
+category = "RiseDev - Start/Stop"
+description = "Run local psql client for postgres with default connection parameters. You can pass extra arguments to psql."
+dependencies = ["check-and-load-risedev-env-file"]
+script = '''
+#!/usr/bin/env bash
+source "${PREFIX_CONFIG}/risedev-env"
+PGHOST=$PGHOST PGPORT=$PGPORT PGUSER=$PGUSER PGDATABASE=$PGDATABASE psql "$@"
+'''
+
 [tasks.ctl]
 category = "RiseDev - Start/Stop"
 description = "Start RiseCtl"
@@ -863,11 +869,6 @@ script = """
 #!/usr/bin/env bash
 set -e
 
-echo "Running Planner Test requires larger stack size, setting RUST_MIN_STACK to 8388608 (8MB) as default."
-if [[ -z "${RUST_MIN_STACK}" ]]; then
-  export RUST_MIN_STACK=8388608
-fi
-
 cargo nextest run --workspace --exclude risingwave_simulation "$@"
 """
 description = "🌟 Run unit tests"
@@ -1208,7 +1209,7 @@ script = """
 
 echo "Running $(tput setaf 4)cargo udeps$(tput sgr0) checks"
 
-cargo udeps --workspace --all-targets ${RISINGWAVE_FEATURE_FLAGS} --exclude workspace-hack --exclude risingwave_bench --exclude risingwave_udf --exclude risingwave_simulation
+cargo udeps --workspace --all-targets ${RISINGWAVE_FEATURE_FLAGS} --exclude workspace-hack --exclude risingwave_bench --exclude risingwave_simulation
 """
 
 [tasks.check-trailing-spaces]
@@ -1303,7 +1304,7 @@ echo If you still feel this is not enough, you may copy $(tput setaf 4)risedev$(
 [tasks.ci-start]
 category = "RiseDev - CI"
 dependencies = ["clean-data", "pre-start-dev"]
-command = "target/debug/risedev-dev"                                                 # `risedev-dev` is always built in dev profile
+command = "target/debug/risedev-dev" # `risedev-dev` is always built in dev profile
 env = { RISEDEV_CLEAN_START = true }
 args = ["${@}"]
 description = "Clean data and start a full RisingWave dev cluster using risedev-dev"
@@ -1340,7 +1341,7 @@ echo "All processes has exited."
 
 [tasks.slt]
 category = "RiseDev - Test - SQLLogicTest"
-install_crate = { min_version = "0.23.1", crate_name = "sqllogictest-bin", binary = "sqllogictest", test_arg = [
+install_crate = { min_version = "0.24.0", crate_name = "sqllogictest-bin", binary = "sqllogictest", test_arg = [
     "--help",
 ], install_command = "binstall" }
 dependencies = ["check-and-load-risedev-env-file"]
diff --git a/ci/Dockerfile b/ci/Dockerfile
index 21b9c30c678b0..7478d24b66edc 100644
--- a/ci/Dockerfile
+++ b/ci/Dockerfile
@@ -70,7 +70,7 @@ ENV CARGO_REGISTRIES_CRATES_IO_PROTOCOL=sparse
 RUN curl -L --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/cargo-bins/cargo-binstall/main/install-from-binstall-release.sh | bash
 RUN cargo binstall -y --locked --no-symlinks cargo-llvm-cov cargo-nextest cargo-hakari cargo-sort cargo-cache cargo-audit \
     cargo-make@0.37.9 \
-    sqllogictest-bin@0.23.1 \
+    sqllogictest-bin@0.24.0 \
     sccache@0.7.4 \
     && cargo cache -a \
     && rm -rf "/root/.cargo/registry/index" \
diff --git a/ci/build-ci-image.sh b/ci/build-ci-image.sh
index 86392d2ecd7d8..969239dea82c2 100755
--- a/ci/build-ci-image.sh
+++ b/ci/build-ci-image.sh
@@ -10,7 +10,7 @@ cat ../rust-toolchain
 # shellcheck disable=SC2155
 
 # REMEMBER TO ALSO UPDATE ci/docker-compose.yml
-export BUILD_ENV_VERSION=v20241213
+export BUILD_ENV_VERSION=v20241224
 
 export BUILD_TAG="public.ecr.aws/w1p7b4n3/rw-build-env:${BUILD_ENV_VERSION}"
 
diff --git a/ci/docker-compose.yml b/ci/docker-compose.yml
index 7ef17777b390c..4b8d795c142f5 100644
--- a/ci/docker-compose.yml
+++ b/ci/docker-compose.yml
@@ -90,7 +90,7 @@ services:
       retries: 5
 
   source-test-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     depends_on:
       - mysql
       - mysql-meta
@@ -106,7 +106,7 @@ services:
       - ..:/risingwave
 
   sink-test-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     depends_on:
       - mysql
       - mysql-meta
@@ -129,13 +129,13 @@ services:
       - ..:/risingwave
 
   rw-build-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     volumes:
       - ..:/risingwave
 
   # Standard environment for CI, including MySQL and Postgres for metadata.
   ci-standard-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     depends_on:
       - mysql-meta
       - db
@@ -143,14 +143,14 @@ services:
       - ..:/risingwave
 
   iceberg-engine-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     depends_on:
       - db
     volumes:
       - ..:/risingwave
 
   ci-flamegraph-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     # NOTE(kwannoel): This is used in order to permit
     # syscalls for `nperf` (perf_event_open),
     # so it can do CPU profiling.
@@ -161,7 +161,7 @@ services:
       - ..:/risingwave
 
   regress-test-env:
-    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241213
+    image: public.ecr.aws/w1p7b4n3/rw-build-env:v20241224
     depends_on:
       db:
         condition: service_healthy
diff --git a/ci/scripts/common.sh b/ci/scripts/common.sh
index b3e09bd607b28..9d807070ca57a 100755
--- a/ci/scripts/common.sh
+++ b/ci/scripts/common.sh
@@ -16,12 +16,13 @@ export GCLOUD_DOWNLOAD_TGZ=https://rw-ci-deps-dist.s3.amazonaws.com/google-cloud
 export NEXTEST_HIDE_PROGRESS_BAR=true
 export RW_TELEMETRY_TYPE=test
 export RW_SECRET_STORE_PRIVATE_KEY_HEX="0123456789abcdef0123456789abcdef"
-export RUST_MIN_STACK=4194304
 
 unset LANG
 
 function dump_diagnose_info() {
-  ./risedev diagnose || true
+  if [ -f .risingwave/config/risedev-env ]; then
+    ./risedev diagnose || true
+  fi
 }
 trap dump_diagnose_info EXIT
 
diff --git a/ci/scripts/e2e-source-test.sh b/ci/scripts/e2e-source-test.sh
index 94ceaaaa853c1..20a06a18de7a4 100755
--- a/ci/scripts/e2e-source-test.sh
+++ b/ci/scripts/e2e-source-test.sh
@@ -144,7 +144,6 @@ risedev ci-kill
 export RISINGWAVE_CI=true
 
 echo "--- e2e, ci-kafka-plus-pubsub, legacy kafka tests"
-export RUST_MIN_STACK=4194304
 RUST_LOG="info,risingwave_stream=info,risingwave_batch=info,risingwave_storage=info" \
 risedev ci-start ci-kafka
 ./e2e_test/source_legacy/basic/scripts/prepare_ci_kafka.sh
diff --git a/ci/scripts/pr-unit-test.sh b/ci/scripts/pr-unit-test.sh
index 73b50909802bf..25f11f446206e 100755
--- a/ci/scripts/pr-unit-test.sh
+++ b/ci/scripts/pr-unit-test.sh
@@ -5,8 +5,4 @@ set -euo pipefail
 
 source ci/scripts/common.sh
 source ci/scripts/pr.env.sh
-
-# Set RUST_MIN_STACK to 8MB to avoid stack overflow in planner test.
-# This is a Unit Test specific setting.
-export RUST_MIN_STACK=8388608
 ./ci/scripts/run-unit-test.sh
diff --git a/dashboard/components/FragmentGraph.tsx b/dashboard/components/FragmentGraph.tsx
index 1e7d80c8641c3..2001cbf79c6fd 100644
--- a/dashboard/components/FragmentGraph.tsx
+++ b/dashboard/components/FragmentGraph.tsx
@@ -24,8 +24,14 @@ import {
   layoutItem,
 } from "../lib/layout"
 import { PlanNodeDatum } from "../pages/fragment_graph"
+import { FragmentStats } from "../proto/gen/monitor_service"
 import { StreamNode } from "../proto/gen/stream_plan"
-import { backPressureColor, backPressureWidth } from "./utils/backPressure"
+import {
+  backPressureColor,
+  backPressureWidth,
+  epochToUnixMillis,
+  latencyToColor,
+} from "./utils/backPressure"
 
 const ReactJson = loadable(() => import("react-json-view"))
 
@@ -95,11 +101,13 @@ export default function FragmentGraph({
   fragmentDependency,
   selectedFragmentId,
   backPressures,
+  fragmentStats,
 }: {
   planNodeDependencies: Map<string, d3.HierarchyNode<PlanNodeDatum>>
   fragmentDependency: FragmentBox[]
   selectedFragmentId?: string
   backPressures?: Map<string, number>
+  fragmentStats?: { [fragmentId: number]: FragmentStats }
 }) {
   const svgRef = useRef<SVGSVGElement>(null)
 
@@ -188,6 +196,7 @@ export default function FragmentGraph({
 
   useEffect(() => {
     if (fragmentLayout) {
+      const now_ms = Date.now()
       const svgNode = svgRef.current
       const svgSelection = d3.select(svgNode)
 
@@ -246,13 +255,69 @@ export default function FragmentGraph({
           .attr("height", ({ height }) => height - fragmentMarginY * 2)
           .attr("x", fragmentMarginX)
           .attr("y", fragmentMarginY)
-          .attr("fill", "white")
+          .attr(
+            "fill",
+            fragmentStats
+              ? ({ id }) => {
+                  const fragmentId = parseInt(id)
+                  if (isNaN(fragmentId) || !fragmentStats[fragmentId]) {
+                    return "white"
+                  }
+                  let currentMs = epochToUnixMillis(
+                    fragmentStats[fragmentId].currentEpoch
+                  )
+                  return latencyToColor(now_ms - currentMs, "white")
+                }
+              : "white"
+          )
           .attr("stroke-width", ({ id }) => (isSelected(id) ? 3 : 1))
           .attr("rx", 5)
           .attr("stroke", ({ id }) =>
             isSelected(id) ? theme.colors.blue[500] : theme.colors.gray[500]
           )
 
+        const getTooltipContent = (id: string) => {
+          const fragmentId = parseInt(id)
+          const stats = fragmentStats?.[fragmentId]
+          const latencySeconds = stats
+            ? ((now_ms - epochToUnixMillis(stats.currentEpoch)) / 1000).toFixed(
+                2
+              )
+            : "N/A"
+          const epoch = stats?.currentEpoch ?? "N/A"
+
+          return `<b>Fragment ${fragmentId}</b><br>Epoch: ${epoch}<br>Latency: ${latencySeconds} seconds`
+        }
+
+        boundingBox
+          .on("mouseover", (event, { id }) => {
+            // Remove existing tooltip if any
+            d3.selectAll(".tooltip").remove()
+
+            // Create new tooltip
+            d3.select("body")
+              .append("div")
+              .attr("class", "tooltip")
+              .style("position", "absolute")
+              .style("background", "white")
+              .style("padding", "10px")
+              .style("border", "1px solid #ddd")
+              .style("border-radius", "4px")
+              .style("pointer-events", "none")
+              .style("left", event.pageX + 10 + "px")
+              .style("top", event.pageY + 10 + "px")
+              .style("font-size", "12px")
+              .html(getTooltipContent(id))
+          })
+          .on("mousemove", (event) => {
+            d3.select(".tooltip")
+              .style("left", event.pageX + 10 + "px")
+              .style("top", event.pageY + 10 + "px")
+          })
+          .on("mouseout", () => {
+            d3.selectAll(".tooltip").remove()
+          })
+
         // Stream node edges
         let edgeSelection = gSel.select<SVGGElement>(".edges")
         if (edgeSelection.empty()) {
@@ -409,24 +474,39 @@ export default function FragmentGraph({
           .attr("stroke-width", width)
           .attr("stroke", color)
 
-        // Tooltip for back pressure rate
-        let title = gSel.select<SVGTitleElement>("title")
-        if (title.empty()) {
-          title = gSel.append<SVGTitleElement>("title")
-        }
-
-        const text = (d: Edge) => {
-          if (backPressures) {
-            let value = backPressures.get(`${d.target}_${d.source}`)
-            if (value) {
-              return `${value.toFixed(2)}%`
+        path
+          .on("mouseover", (event, d) => {
+            // Remove existing tooltip if any
+            d3.selectAll(".tooltip").remove()
+
+            if (backPressures) {
+              const value = backPressures.get(`${d.target}_${d.source}`)
+              if (value) {
+                // Create new tooltip
+                d3.select("body")
+                  .append("div")
+                  .attr("class", "tooltip")
+                  .style("position", "absolute")
+                  .style("background", "white")
+                  .style("padding", "10px")
+                  .style("border", "1px solid #ddd")
+                  .style("border-radius", "4px")
+                  .style("pointer-events", "none")
+                  .style("left", event.pageX + 10 + "px")
+                  .style("top", event.pageY + 10 + "px")
+                  .style("font-size", "12px")
+                  .html(`BP: ${value.toFixed(2)}%`)
+              }
             }
-          }
-
-          return ""
-        }
-
-        title.text(text)
+          })
+          .on("mousemove", (event) => {
+            d3.select(".tooltip")
+              .style("left", event.pageX + 10 + "px")
+              .style("top", event.pageY + 10 + "px")
+          })
+          .on("mouseout", () => {
+            d3.selectAll(".tooltip").remove()
+          })
 
         return gSel
       }
diff --git a/dashboard/components/RelationGraph.tsx b/dashboard/components/RelationGraph.tsx
index a61a6924a370e..59c87275b43ed 100644
--- a/dashboard/components/RelationGraph.tsx
+++ b/dashboard/components/RelationGraph.tsx
@@ -33,8 +33,14 @@ import {
   flipLayoutRelation,
   generateRelationEdges,
 } from "../lib/layout"
+import { RelationStats } from "../proto/gen/monitor_service"
 import { CatalogModal, useCatalogModal } from "./CatalogModal"
-import { backPressureColor, backPressureWidth } from "./utils/backPressure"
+import {
+  backPressureColor,
+  backPressureWidth,
+  epochToUnixMillis,
+  latencyToColor,
+} from "./utils/backPressure"
 
 function boundBox(
   relationPosition: RelationPointPosition[],
@@ -62,11 +68,13 @@ export default function RelationGraph({
   selectedId,
   setSelectedId,
   backPressures,
+  relationStats,
 }: {
   nodes: RelationPoint[]
   selectedId: string | undefined
   setSelectedId: (id: string) => void
   backPressures?: Map<string, number> // relationId-relationId->back_pressure_rate})
+  relationStats: { [relationId: number]: RelationStats } | undefined
 }) {
   const [modalData, setModalId] = useCatalogModal(nodes.map((n) => n.relation))
 
@@ -99,6 +107,7 @@ export default function RelationGraph({
   const { layoutMap, width, height, links } = layoutMapCallback()
 
   useEffect(() => {
+    const now_ms = Date.now()
     const svgNode = svgRef.current
     const svgSelection = d3.select(svgNode)
 
@@ -150,24 +159,39 @@ export default function RelationGraph({
           isSelected(d.source) || isSelected(d.target) ? 1 : 0.5
         )
 
-      // Tooltip for back pressure rate
-      let title = sel.select<SVGTitleElement>("title")
-      if (title.empty()) {
-        title = sel.append<SVGTitleElement>("title")
-      }
-
-      const text = (d: Edge) => {
-        if (backPressures) {
-          let value = backPressures.get(`${d.target}_${d.source}`)
-          if (value) {
-            return `${value.toFixed(2)}%`
+      sel
+        .on("mouseover", (event, d) => {
+          // Remove existing tooltip if any
+          d3.selectAll(".tooltip").remove()
+
+          if (backPressures) {
+            const value = backPressures.get(`${d.target}_${d.source}`)
+            if (value) {
+              // Create new tooltip
+              d3.select("body")
+                .append("div")
+                .attr("class", "tooltip")
+                .style("position", "absolute")
+                .style("background", "white")
+                .style("padding", "10px")
+                .style("border", "1px solid #ddd")
+                .style("border-radius", "4px")
+                .style("pointer-events", "none")
+                .style("left", event.pageX + 10 + "px")
+                .style("top", event.pageY + 10 + "px")
+                .style("font-size", "12px")
+                .html(`BP: ${value.toFixed(2)}%`)
+            }
           }
-        }
-
-        return ""
-      }
-
-      title.text(text)
+        })
+        .on("mousemove", (event) => {
+          d3.select(".tooltip")
+            .style("left", event.pageX + 10 + "px")
+            .style("top", event.pageY + 10 + "px")
+        })
+        .on("mouseout", () => {
+          d3.selectAll(".tooltip").remove()
+        })
 
       return sel
     }
@@ -189,9 +213,19 @@ export default function RelationGraph({
 
       circle.attr("r", nodeRadius).attr("fill", ({ id, relation }) => {
         const weight = relationIsStreamingJob(relation) ? "500" : "400"
-        return isSelected(id)
+        const baseColor = isSelected(id)
           ? theme.colors.blue[weight]
           : theme.colors.gray[weight]
+        if (relationStats) {
+          const relationId = parseInt(id)
+          if (!isNaN(relationId) && relationStats[relationId]) {
+            const currentMs = epochToUnixMillis(
+              relationStats[relationId].currentEpoch
+            )
+            return latencyToColor(now_ms - currentMs, baseColor)
+          }
+        }
+        return baseColor
       })
 
       // Relation name
@@ -233,16 +267,50 @@ export default function RelationGraph({
         .attr("font-size", 16)
         .attr("font-weight", "bold")
 
-      // Relation type tooltip
-      let typeTooltip = g.select<SVGTitleElement>("title")
-      if (typeTooltip.empty()) {
-        typeTooltip = g.append<SVGTitleElement>("title")
+      // Tooltip
+      const getTooltipContent = (relation: Relation, id: string) => {
+        const relationId = parseInt(id)
+        const stats = relationStats?.[relationId]
+        const latencySeconds = stats
+          ? (
+              (Date.now() - epochToUnixMillis(stats.currentEpoch)) /
+              1000
+            ).toFixed(2)
+          : "N/A"
+        const epoch = stats?.currentEpoch ?? "N/A"
+
+        return `<b>${relation.name} (${relationTypeTitleCase(
+          relation
+        )})</b><br>Epoch: ${epoch}<br>Latency: ${latencySeconds} seconds`
       }
 
-      typeTooltip.text(
-        ({ relation }) =>
-          `${relation.name} (${relationTypeTitleCase(relation)})`
-      )
+      g.on("mouseover", (event, { relation, id }) => {
+        // Remove existing tooltip if any
+        d3.selectAll(".tooltip").remove()
+
+        // Create new tooltip
+        d3.select("body")
+          .append("div")
+          .attr("class", "tooltip")
+          .style("position", "absolute")
+          .style("background", "white")
+          .style("padding", "10px")
+          .style("border", "1px solid #ddd")
+          .style("border-radius", "4px")
+          .style("pointer-events", "none")
+          .style("left", event.pageX + 10 + "px")
+          .style("top", event.pageY + 10 + "px")
+          .style("font-size", "12px")
+          .html(getTooltipContent(relation, id))
+      })
+        .on("mousemove", (event) => {
+          d3.select(".tooltip")
+            .style("left", event.pageX + 10 + "px")
+            .style("top", event.pageY + 10 + "px")
+        })
+        .on("mouseout", () => {
+          d3.selectAll(".tooltip").remove()
+        })
 
       // Relation modal
       g.style("cursor", "pointer").on("click", (_, { relation, id }) => {
@@ -265,7 +333,15 @@ export default function RelationGraph({
     nodeSelection.enter().call(createNode)
     nodeSelection.call(applyNode)
     nodeSelection.exit().remove()
-  }, [layoutMap, links, selectedId, setModalId, setSelectedId, backPressures])
+  }, [
+    layoutMap,
+    links,
+    selectedId,
+    setModalId,
+    setSelectedId,
+    backPressures,
+    relationStats,
+  ])
 
   return (
     <>
diff --git a/dashboard/components/utils/backPressure.tsx b/dashboard/components/utils/backPressure.tsx
index afb26e0746da4..841e64baafa82 100644
--- a/dashboard/components/utils/backPressure.tsx
+++ b/dashboard/components/utils/backPressure.tsx
@@ -41,3 +41,47 @@ export function backPressureWidth(value: number, scale: number) {
 
   return scale * (value / 100) + 2
 }
+
+export function epochToUnixMillis(epoch: number) {
+  // UNIX_RISINGWAVE_DATE_SEC
+  return 1617235200000 + epoch / 65536
+}
+
+export function latencyToColor(latency_ms: number, baseColor: string) {
+  const LOWER = 10000 // 10s
+  const UPPER = 300000 // 5min
+
+  const colorRange = [
+    baseColor,
+    theme.colors.yellow["200"],
+    theme.colors.orange["300"],
+    theme.colors.red["400"],
+  ].map((c) => tinycolor(c))
+
+  if (latency_ms <= LOWER) {
+    return baseColor
+  }
+
+  if (latency_ms >= UPPER) {
+    return theme.colors.red["400"]
+  }
+
+  // Map log(latency) to [0,1] range between 10s and 5min
+  const minLog = Math.log(LOWER)
+  const maxLog = Math.log(UPPER)
+  const latencyLog = Math.log(latency_ms)
+  const normalizedPos = (latencyLog - minLog) / (maxLog - minLog)
+
+  // Map to color range index
+  const step = colorRange.length - 1
+  const pos = normalizedPos * step
+  const floor = Math.floor(pos)
+  const ceil = Math.ceil(pos)
+
+  // Interpolate between colors
+  const color = tinycolor(colorRange[floor])
+    .mix(tinycolor(colorRange[ceil]), (pos - floor) * 100)
+    .toHexString()
+
+  return color
+}
diff --git a/dashboard/lib/api/api.ts b/dashboard/lib/api/api.ts
index af0116e31f3bb..48af21bc4c620 100644
--- a/dashboard/lib/api/api.ts
+++ b/dashboard/lib/api/api.ts
@@ -26,7 +26,9 @@ export const PREDEFINED_API_ENDPOINTS = [
 ]
 
 export const DEFAULT_API_ENDPOINT: string =
-  process.env.NODE_ENV === "production" ? PROD_API_ENDPOINT : MOCK_API_ENDPOINT // EXTERNAL_META_NODE_API_ENDPOINT to debug with RisingWave servers
+  process.env.NODE_ENV === "production"
+    ? PROD_API_ENDPOINT
+    : EXTERNAL_META_NODE_API_ENDPOINT // EXTERNAL_META_NODE_API_ENDPOINT to debug with RisingWave servers
 
 export const API_ENDPOINT_KEY = "risingwave.dashboard.api.endpoint"
 
diff --git a/dashboard/lib/api/metric.ts b/dashboard/lib/api/metric.ts
index f310e0572170d..23d007bb08360 100644
--- a/dashboard/lib/api/metric.ts
+++ b/dashboard/lib/api/metric.ts
@@ -156,9 +156,7 @@ export async function fetchEmbeddedBackPressure() {
   const response: GetBackPressureResponse = await api.get(
     "/metrics/fragment/embedded_back_pressures"
   )
-  let backPressureInfos: BackPressureInfo[] =
-    response.backPressureInfos?.map(BackPressureInfo.fromJSON) ?? []
-  return backPressureInfos
+  return response
 }
 
 function calculatePercentile(samples: MetricsSample[], percentile: number) {
diff --git a/dashboard/pages/fragment_graph.tsx b/dashboard/pages/fragment_graph.tsx
index 2312334e53abe..4b2a39aa5585c 100644
--- a/dashboard/pages/fragment_graph.tsx
+++ b/dashboard/pages/fragment_graph.tsx
@@ -56,7 +56,7 @@ import {
 } from "../lib/api/streaming"
 import { FragmentBox } from "../lib/layout"
 import { TableFragments, TableFragments_Fragment } from "../proto/gen/meta"
-import { BackPressureInfo } from "../proto/gen/monitor_service"
+import { BackPressureInfo, FragmentStats } from "../proto/gen/monitor_service"
 import { Dispatcher, MergeNode, StreamNode } from "../proto/gen/stream_plan"
 
 interface DispatcherNode {
@@ -329,12 +329,18 @@ export default function Streaming() {
   // Didn't call `useFetch()` because the `setState` way is special.
   const [embeddedBackPressureInfo, setEmbeddedBackPressureInfo] =
     useState<EmbeddedBackPressureInfo>()
+  const [fragmentStats, setFragmentStats] = useState<{
+    [key: number]: FragmentStats
+  }>()
+
   useEffect(() => {
     if (backPressureDataSource === "Embedded") {
-      const interval = setInterval(() => {
+      function refresh() {
         fetchEmbeddedBackPressure().then(
-          (newBP) => {
-            console.log(newBP)
+          (response) => {
+            console.log(response)
+            let newBP =
+              response.backPressureInfos?.map(BackPressureInfo.fromJSON) ?? []
             setEmbeddedBackPressureInfo((prev) =>
               prev
                 ? {
@@ -355,13 +361,16 @@ export default function Streaming() {
                     totalDurationNs: 0,
                   }
             )
+            setFragmentStats(response.fragmentStats)
           },
           (e) => {
             console.error(e)
             toast(e, "error")
           }
         )
-      }, INTERVAL_MS)
+      }
+      refresh()
+      const interval = setInterval(refresh, INTERVAL_MS)
       return () => {
         clearInterval(interval)
       }
@@ -550,6 +559,7 @@ export default function Streaming() {
               fragmentDependency={fragmentDependency}
               planNodeDependencies={planNodeDependencies}
               backPressures={backPressures}
+              fragmentStats={fragmentStats}
             />
           )}
         </Box>
diff --git a/dashboard/pages/relation_graph.tsx b/dashboard/pages/relation_graph.tsx
index a1061ffa23f0d..d44b29708bdc4 100644
--- a/dashboard/pages/relation_graph.tsx
+++ b/dashboard/pages/relation_graph.tsx
@@ -47,7 +47,7 @@ import {
   relationIsStreamingJob,
 } from "../lib/api/streaming"
 import { RelationPoint } from "../lib/layout"
-import { BackPressureInfo } from "../proto/gen/monitor_service"
+import { BackPressureInfo, RelationStats } from "../proto/gen/monitor_service"
 
 const SIDEBAR_WIDTH = "200px"
 const INTERVAL_MS = 5000
@@ -133,15 +133,20 @@ export default function StreamingGraph() {
   // Didn't call `useFetch()` because the `setState` way is special.
   const [embeddedBackPressureInfo, setEmbeddedBackPressureInfo] =
     useState<EmbeddedBackPressureInfo>()
+  const [relationStats, setRelationStats] = useState<{
+    [key: number]: RelationStats
+  }>()
+
   useEffect(() => {
     if (resetEmbeddedBackPressures) {
       setEmbeddedBackPressureInfo(undefined)
       toggleResetEmbeddedBackPressures()
     }
     if (backPressureDataSource === "Embedded") {
-      const interval = setInterval(() => {
+      function refresh() {
         fetchEmbeddedBackPressure().then(
-          (newBP) => {
+          (response) => {
+            let newBP = response.backPressureInfos
             setEmbeddedBackPressureInfo((prev) =>
               prev
                 ? {
@@ -162,13 +167,16 @@ export default function StreamingGraph() {
                     totalDurationNs: 0,
                   }
             )
+            setRelationStats(response.relationStats)
           },
           (e) => {
             console.error(e)
             toast(e, "error")
           }
         )
-      }, INTERVAL_MS)
+      }
+      refresh()
+      const interval = setInterval(refresh, INTERVAL_MS)
       return () => {
         clearInterval(interval)
       }
@@ -308,6 +316,7 @@ export default function StreamingGraph() {
               selectedId={selectedId?.toString()}
               setSelectedId={(id) => setSelectedId(parseInt(id))}
               backPressures={backPressures}
+              relationStats={relationStats}
             />
           )}
         </Box>
diff --git a/docker/docker-compose-distributed.yml b/docker/docker-compose-distributed.yml
index 2ad66049f1a98..a4d5c8760bbe2 100644
--- a/docker/docker-compose-distributed.yml
+++ b/docker/docker-compose-distributed.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   compactor-0:
     <<: *image
diff --git a/docker/docker-compose-with-azblob.yml b/docker/docker-compose-with-azblob.yml
index 6b32b7357c8a3..ec74ea64a5210 100644
--- a/docker/docker-compose-with-azblob.yml
+++ b/docker/docker-compose-with-azblob.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose-with-gcs.yml b/docker/docker-compose-with-gcs.yml
index 7a25348a2c9ff..d66e60af4330d 100644
--- a/docker/docker-compose-with-gcs.yml
+++ b/docker/docker-compose-with-gcs.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose-with-local-fs.yml b/docker/docker-compose-with-local-fs.yml
index 9bfd551809328..fece237915243 100644
--- a/docker/docker-compose-with-local-fs.yml
+++ b/docker/docker-compose-with-local-fs.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose-with-obs.yml b/docker/docker-compose-with-obs.yml
index e1665b0ee0298..41fbd75c852a5 100644
--- a/docker/docker-compose-with-obs.yml
+++ b/docker/docker-compose-with-obs.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose-with-oss.yml b/docker/docker-compose-with-oss.yml
index de9e4c727e760..574a099e0dedc 100644
--- a/docker/docker-compose-with-oss.yml
+++ b/docker/docker-compose-with-oss.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose-with-s3.yml b/docker/docker-compose-with-s3.yml
index b00149c0c175b..71672ca6f21a8 100644
--- a/docker/docker-compose-with-s3.yml
+++ b/docker/docker-compose-with-s3.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose-with-sqlite.yml b/docker/docker-compose-with-sqlite.yml
index ae0b2f31b07b2..0e1e228cfee91 100644
--- a/docker/docker-compose-with-sqlite.yml
+++ b/docker/docker-compose-with-sqlite.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/docker/docker-compose.yml b/docker/docker-compose.yml
index eb0ee6a9b02a2..e7ab7a794e3f5 100644
--- a/docker/docker-compose.yml
+++ b/docker/docker-compose.yml
@@ -1,6 +1,6 @@
 ---
 x-image: &image
-  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0-rc.2}
+  image: ${RW_IMAGE:-risingwavelabs/risingwave:v2.1.0}
 services:
   risingwave-standalone:
     <<: *image
diff --git a/e2e_test/batch/basic/row_filter.slt.part b/e2e_test/batch/basic/row_filter.slt.part
new file mode 100644
index 0000000000000..fdec809d92379
--- /dev/null
+++ b/e2e_test/batch/basic/row_filter.slt.part
@@ -0,0 +1,65 @@
+statement ok
+create table t1(v1 int, v2 int, v3 int);
+
+statement ok
+insert into t1 values(1,1,1),(1,2,1),(1,2,2),(1,3,1),(1,3,2),(1,3,3);
+
+statement ok
+create materialized view mv1 as select * from t1 order by v1 asc, v2 desc, v3 asc;
+
+statement ok
+create materialized view mv2 as select * from t1 order by v1 desc, v2 desc, v3 desc;
+
+statement ok
+create materialized view mv3 as select * from t1 order by v1 asc, v2 asc, v3 asc;
+
+query III
+select * from mv1 where (v1,v2,v3) > (1,3,1) order by v3;
+----
+1 3 2
+1 3 3
+
+query III
+select * from mv2 where (v1,v2,v3) > (1,3,1) order by v3;
+----
+1 3 2
+1 3 3
+
+query III
+select * from mv3 where (v1,v2,v3) > (1,3,1) order by v3;
+----
+1 3 2
+1 3 3
+
+query III
+select * from mv1 where (v1,v2,v3) < (1,3,1) order by v1,v2,v3;
+----
+1 1 1
+1 2 1
+1 2 2
+
+query III
+select * from mv2 where (v1,v2,v3) < (1,3,1) order by v1,v2,v3;
+----
+1 1 1
+1 2 1
+1 2 2
+
+query III
+select * from mv3 where (v1,v2,v3) < (1,3,1) order by v1,v2,v3;
+----
+1 1 1
+1 2 1
+1 2 2
+
+statement ok
+drop materialized view mv3;
+
+statement ok
+drop materialized view mv2;
+
+statement ok
+drop materialized view mv1;
+
+statement ok
+drop table t1;
diff --git a/e2e_test/batch/catalog/information_schema.slt b/e2e_test/batch/catalog/information_schema.slt
new file mode 100644
index 0000000000000..eed5fbd9d68dc
--- /dev/null
+++ b/e2e_test/batch/catalog/information_schema.slt
@@ -0,0 +1,63 @@
+statement ok
+create table t(a int, b bigint, key int primary key);
+
+statement ok
+create view v as select * from t;
+
+statement ok
+create materialized view mv as select * from t;
+
+query TT
+select table_schema, table_name from information_schema.views where table_schema = 'public';
+----
+public v
+
+query TT
+select table_schema, table_name from information_schema.tables where table_schema = 'public' order by table_name;
+----
+public mv
+public t
+public v
+
+query TTTTTTTT
+select constraint_schema, constraint_name, table_schema, table_name from information_schema.table_constraints where table_schema = 'public' order by table_name;
+----
+public mv_pkey public mv
+public t_pkey public t
+
+query TT
+select schema_name from information_schema.schemata order by schema_name;
+----
+information_schema
+pg_catalog
+public
+rw_catalog
+
+query TTTTTII
+select * EXCEPT(constraint_catalog, table_catalog) from information_schema.key_column_usage where table_schema = 'public' order by table_name;
+----
+public mv_pkey public mv key 3 NULL
+public t_pkey public t key 3 NULL
+
+query TTTITTT
+select table_schema, table_name, column_name, ordinal_position, data_type, udt_schema, udt_name from information_schema.columns where table_schema = 'public' order by table_name, ordinal_position;
+----
+public mv a 1 integer pg_catalog int4
+public mv b 2 bigint pg_catalog int8
+public mv key 3 integer pg_catalog int4
+public t a 1 integer pg_catalog int4
+public t b 2 bigint pg_catalog int8
+public t key 3 integer pg_catalog int4
+public v a 1 integer pg_catalog int4
+public v b 2 bigint pg_catalog int8
+public v key 3 integer pg_catalog int4
+
+
+statement ok
+drop materialized view mv;
+
+statement ok
+drop view v;
+
+statement ok
+drop table t;
diff --git a/e2e_test/batch/catalog/pg_class.slt.part b/e2e_test/batch/catalog/pg_class.slt.part
index f15a2c74fdb40..14aa5c85c2123 100644
--- a/e2e_test/batch/catalog/pg_class.slt.part
+++ b/e2e_test/batch/catalog/pg_class.slt.part
@@ -2,22 +2,22 @@ query ITIT
 SELECT oid,relname,relowner,relkind FROM pg_catalog.pg_class ORDER BY oid limit 15;
 ----
 2147478647 columns 1 v
-2147478648 schemata 1 v
-2147478649 table_constraints 1 v
-2147478650 tables 1 v
-2147478651 views 1 v
-2147478652 pg_am 1 v
-2147478653 pg_attrdef 1 v
-2147478654 pg_attribute 1 v
-2147478655 pg_auth_members 1 v
-2147478656 pg_cast 1 r
-2147478657 pg_class 1 v
-2147478658 pg_collation 1 v
-2147478659 pg_constraint 1 r
-2147478660 pg_conversion 1 v
-2147478661 pg_database 1 v
+2147478648 key_column_usage 1 v
+2147478649 schemata 1 v
+2147478650 table_constraints 1 v
+2147478651 tables 1 v
+2147478652 views 1 v
+2147478653 pg_am 1 v
+2147478654 pg_attrdef 1 v
+2147478655 pg_attribute 1 v
+2147478656 pg_auth_members 1 v
+2147478657 pg_cast 1 r
+2147478658 pg_class 1 v
+2147478659 pg_collation 1 v
+2147478660 pg_constraint 1 r
+2147478661 pg_conversion 1 v
 
 query ITIT
 SELECT oid,relname,relowner,relkind FROM pg_catalog.pg_class WHERE oid = 'pg_namespace'::regclass;
 ----
-2147478672 pg_namespace 1 v
+2147478673 pg_namespace 1 v
diff --git a/e2e_test/sink/pg_native_vs_jdbc.slt b/e2e_test/sink/pg_native_vs_jdbc.slt
new file mode 100644
index 0000000000000..288d3cdc41053
--- /dev/null
+++ b/e2e_test/sink/pg_native_vs_jdbc.slt
@@ -0,0 +1,84 @@
+control substitution on
+
+statement ok
+DROP SINK IF EXISTS pg_sink;
+
+statement ok
+DROP SINK IF EXISTS pg_sink_jdbc;
+
+statement ok
+DROP TABLE IF EXISTS datagen_source;
+
+system ok
+psql -c 'DROP TABLE IF EXISTS datagen_source;'
+
+system ok
+psql -c 'DROP TABLE IF EXISTS datagen_source_jdbc;'
+
+system ok
+psql -c 'CREATE TABLE datagen_source (id INT PRIMARY KEY, v1 varchar);'
+
+system ok
+psql -c 'CREATE TABLE datagen_source_jdbc (id INT PRIMARY KEY, v1 varchar);'
+
+statement ok
+CREATE TABLE datagen_source (id INT PRIMARY KEY, v1 varchar);
+
+statement ok
+INSERT INTO datagen_source SELECT key, 'asjdkk2kbdk2uk2ubek2' FROM generate_series(1, 2000000) t(key);
+
+statement ok
+flush;
+
+statement ok
+CREATE SINK pg_sink FROM datagen_source WITH (
+    connector='postgres',
+    host='$PGHOST',
+    port='$PGPORT',
+    user='$PGUSER',
+    password='$PGPASSWORD',
+    database='$PGDATABASE',
+    table='datagen_source',
+    type='upsert',
+    primary_key='id',
+);
+
+sleep 240s
+
+system ok
+psql --tuples-only -c 'select count(*) from datagen_source;'
+----
+ 2000000
+
+
+statement ok
+CREATE SINK pg_sink_jdbc FROM datagen_source WITH (
+    connector='jdbc',
+    jdbc.url='jdbc:postgresql://${PGHOST}:${PGPORT}/${PGDATABASE}?user=${PGUSER}&password=${PGPASSWORD}',
+    table.name='datagen_source_jdbc',
+    primary_key='id',
+    type='upsert'
+);
+
+sleep 240s
+
+system ok
+psql --tuples-only -c 'select count(*) from datagen_source_jdbc;'
+----
+ 2000000
+
+
+statement ok
+DROP SINK IF EXISTS pg_sink;
+
+statement ok
+DROP SINK IF EXISTS pg_sink_jdbc;
+
+statement ok
+DROP TABLE IF EXISTS datagen_source;
+
+system ok
+psql -c 'DROP TABLE IF EXISTS datagen_source;'
+
+system ok
+psql -c 'DROP TABLE IF EXISTS datagen_source_jdbc;'
\ No newline at end of file
diff --git a/integration_tests/feature-store/server/Cargo.toml b/integration_tests/feature-store/server/Cargo.toml
index 26e300cdf2f6d..8e13d5374d7bf 100644
--- a/integration_tests/feature-store/server/Cargo.toml
+++ b/integration_tests/feature-store/server/Cargo.toml
@@ -9,16 +9,12 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-sqlx = { version = "0.8", features = ["runtime-tokio-native-tls", "postgres"] }
 tokio = { version = "1", features = ["full"] }
 tonic = "0.11.0"
-reqwest = { version = "0.11", features = ["blocking"] }
 rdkafka = { version = "0.34", features = ["cmake-build"] }
-serde_json = "1.0"
 prost = "0.12"
 clap = "4.4.6"
 tokio-postgres = "0.7.10"
-tonic-build = "0.11.0"
 
 [build-dependencies]
 tonic-build = "0.11.0"
diff --git a/integration_tests/feature-store/simulator/Cargo.toml b/integration_tests/feature-store/simulator/Cargo.toml
index e7b53c9047de2..0591ddf5b3137 100644
--- a/integration_tests/feature-store/simulator/Cargo.toml
+++ b/integration_tests/feature-store/simulator/Cargo.toml
@@ -11,12 +11,10 @@ edition = "2021"
 [dependencies]
 tokio = { version = "1", features = ["macros", "rt", "rt-multi-thread"] }
 tonic = "0.11.0"
-reqwest = { version = "0.11" }
 serde_json = "1.0"
 serde_derive = "1.0"
 rand = { workspace = true }
 clap = "4.4.6"
 prost = "0.12"
-serde = { version = "1", features = ["derive"] }
 futures = "0.3.28"
 csv = "1.3.0"
diff --git a/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterHandler.java b/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterHandler.java
index 89bd1dd7e6865..604463e578985 100644
--- a/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterHandler.java
+++ b/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterHandler.java
@@ -44,6 +44,6 @@ public static void runJniSinkWriterThread(long requestRxPtr, long responseTxPtr)
         } catch (Throwable t) {
             sinkWriterStreamObserver.onError(t);
         }
-        LOG.info("end of runJniSinkWriterThread");
+        LOG.debug("end of runJniSinkWriterThread");
     }
 }
diff --git a/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterResponseObserver.java b/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterResponseObserver.java
index 1f1c68f54551e..505a22947b655 100644
--- a/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterResponseObserver.java
+++ b/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/JniSinkWriterResponseObserver.java
@@ -51,7 +51,7 @@ public void onError(Throwable throwable) {
 
     @Override
     public void onCompleted() {
-        LOG.info("JniSinkWriterHandler onCompleted");
+        LOG.debug("JniSinkWriterHandler onCompleted");
     }
 
     public boolean isSuccess() {
diff --git a/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/SinkWriterStreamObserver.java b/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/SinkWriterStreamObserver.java
index 53dfe326fbd9d..55e5befa62ea6 100644
--- a/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/SinkWriterStreamObserver.java
+++ b/java/connector-node/risingwave-connector-service/src/main/java/com/risingwave/connector/SinkWriterStreamObserver.java
@@ -187,7 +187,7 @@ public void onError(Throwable throwable) {
 
     @Override
     public void onCompleted() {
-        LOG.info("sink writer completed");
+        LOG.debug("sink writer completed");
         cleanup();
         responseObserver.onCompleted();
     }
diff --git a/proto/monitor_service.proto b/proto/monitor_service.proto
index 772d43e5613c2..08fc97bb73f9b 100644
--- a/proto/monitor_service.proto
+++ b/proto/monitor_service.proto
@@ -61,8 +61,20 @@ message BackPressureInfo {
   double value = 4;
 }
 
+message FragmentStats {
+  uint32 actor_count = 2;
+  uint64 current_epoch = 3;
+}
+
+message RelationStats {
+  uint32 actor_count = 2;
+  uint64 current_epoch = 3;
+}
+
 message GetBackPressureResponse {
   repeated BackPressureInfo back_pressure_infos = 1;
+  map<uint32, FragmentStats> fragment_stats = 2;
+  map<uint32, RelationStats> relation_stats = 3;
 }
 
 message TieredCacheTracingRequest {
diff --git a/proto/stream_plan.proto b/proto/stream_plan.proto
index 8aabadfd2b74d..3963e3d03e4c3 100644
--- a/proto/stream_plan.proto
+++ b/proto/stream_plan.proto
@@ -627,6 +627,8 @@ message StreamScanNode {
   // The state table used by ArrangementBackfill to replicate upstream mview's state table.
   // Used iff `ChainType::ArrangementBackfill`.
   catalog.Table arrangement_table = 10;
+
+  optional uint64 snapshot_backfill_epoch = 11;
 }
 
 // Config options for CDC backfill
diff --git a/src/batch/Cargo.toml b/src/batch/Cargo.toml
index c501f2df4962e..56491055cfacd 100644
--- a/src/batch/Cargo.toml
+++ b/src/batch/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 async-recursion = "1"
@@ -62,11 +56,9 @@ twox-hash = "2"
 workspace-hack = { path = "../workspace-hack" }
 
 [dev-dependencies]
-criterion = { workspace = true, features = ["async_tokio", "async"] }
 rand = { workspace = true }
 risingwave_hummock_sdk = { workspace = true }
 tempfile = "3"
-tikv-jemallocator = { workspace = true }
 
 [lints]
 workspace = true
diff --git a/src/batch/executors/Cargo.toml b/src/batch/executors/Cargo.toml
index 927c026e83d8e..736c54f8a5b0e 100644
--- a/src/batch/executors/Cargo.toml
+++ b/src/batch/executors/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 assert_matches = "1"
diff --git a/src/batch/executors/src/executor/row_seq_scan.rs b/src/batch/executors/src/executor/row_seq_scan.rs
index 445b1fa8038d7..a9efb465df187 100644
--- a/src/batch/executors/src/executor/row_seq_scan.rs
+++ b/src/batch/executors/src/executor/row_seq_scan.rs
@@ -434,6 +434,7 @@ impl<S: StateStore> RowSeqScanExecutor<S> {
             next_col_bounds,
         } = scan_range;
 
+        // The len of a valid pk_prefix should be less than or equal pk's num.
         let order_type = table.pk_serializer().get_order_types()[pk_prefix.len()];
         let (start_bound, end_bound) = if order_type.is_ascending() {
             (next_col_bounds.0, next_col_bounds.1)
diff --git a/src/cmd/Cargo.toml b/src/cmd/Cargo.toml
index 5ddc6112bd4ab..097f9fd881a48 100644
--- a/src/cmd/Cargo.toml
+++ b/src/cmd/Cargo.toml
@@ -21,7 +21,6 @@ normal = ["workspace-hack", "workspace-config", "task_stats_alloc"]
 
 [dependencies]
 clap = { workspace = true }
-prometheus = { version = "0.13" }
 risingwave_batch_executors = { workspace = true }
 risingwave_common = { workspace = true }
 risingwave_compactor = { workspace = true }
diff --git a/src/cmd_all/Cargo.toml b/src/cmd_all/Cargo.toml
index 2269ca1010481..0e74a20a4a10e 100644
--- a/src/cmd_all/Cargo.toml
+++ b/src/cmd_all/Cargo.toml
@@ -19,18 +19,16 @@ js-udf = ["risingwave_expr_impl/js-udf"]
 python-udf = ["risingwave_expr_impl/python-udf"]
 
 [package.metadata.cargo-machete]
-ignored = ["workspace-hack", "workspace-config", "task_stats_alloc"]
+ignored = ["workspace-hack", "workspace-config", "task_stats_alloc", "tikv-jemallocator"]
 
 [package.metadata.cargo-udeps.ignore]
-ignored = ["workspace-hack", "workspace-config", "task_stats_alloc"]
+ignored = ["workspace-hack", "workspace-config", "task_stats_alloc", "tikv-jemallocator"]
 
 [dependencies]
-anyhow = "1"
 clap = { workspace = true }
 console = "0.15"
 const-str = "0.5"
 home = "0.5"
-prometheus = { version = "0.13" }
 risingwave_batch_executors = { workspace = true }
 risingwave_cmd = { workspace = true }
 risingwave_common = { workspace = true }
diff --git a/src/common/Cargo.toml b/src/common/Cargo.toml
index 8a554137dccdd..1fbf8e312185d 100644
--- a/src/common/Cargo.toml
+++ b/src/common/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 ahash = "0.8"
 anyhow = "1"
@@ -48,12 +42,11 @@ ethnum = { version = "1", features = ["serde"] }
 fixedbitset = "0.5"
 foyer = { workspace = true }
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
-governor = { version = "0.6", default-features = false, features = ["std"] }
+governor = { workspace = true }
 hashbrown = "0.14"
 hex = "0.4.3"
 http = "1"
 humantime = "2.1"
-hytra = { workspace = true }
 itertools = { workspace = true }
 itoa = "1.0"
 jsonbb = { workspace = true }
diff --git a/src/common/common_service/Cargo.toml b/src/common/common_service/Cargo.toml
index 37775ff04a82b..38dd8f4534477 100644
--- a/src/common/common_service/Cargo.toml
+++ b/src/common/common_service/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 async-trait = "0.1"
 axum = { workspace = true }
diff --git a/src/common/estimate_size/Cargo.toml b/src/common/estimate_size/Cargo.toml
index 77e4203f9c7cb..c6447ff974216 100644
--- a/src/common/estimate_size/Cargo.toml
+++ b/src/common/estimate_size/Cargo.toml
@@ -8,19 +8,12 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 bytes = "1"
 educe = "0.6"
 ethnum = { version = "1", features = ["serde"] }
 fixedbitset = "0.5"
 jsonbb = { workspace = true }
-lru = { workspace = true }
 risingwave_common_proc_macro = { workspace = true }
 rust_decimal = "1"
 serde_json = "1"
diff --git a/src/common/heap_profiling/Cargo.toml b/src/common/heap_profiling/Cargo.toml
index 706f801fe4762..edb0d11e2fd0b 100644
--- a/src/common/heap_profiling/Cargo.toml
+++ b/src/common/heap_profiling/Cargo.toml
@@ -8,14 +8,7 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
-anyhow = "1"
 chrono = { version = "0.4", default-features = false, features = [
     "clock",
     "std",
diff --git a/src/common/metrics/Cargo.toml b/src/common/metrics/Cargo.toml
index 78562624c6935..b499508f177d9 100644
--- a/src/common/metrics/Cargo.toml
+++ b/src/common/metrics/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 auto_impl = "1"
 bytes = "1"
diff --git a/src/common/metrics/src/gauge_ext.rs b/src/common/metrics/src/gauge_ext.rs
index b76f2ac7531d0..aa5471ec87171 100644
--- a/src/common/metrics/src/gauge_ext.rs
+++ b/src/common/metrics/src/gauge_ext.rs
@@ -12,8 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use prometheus::core::{AtomicU64, GenericGauge};
 use prometheus::IntGauge;
 
+/// The integer version of [`prometheus::Gauge`]. Provides better performance if metric values are
+/// all unsigned integers.
+pub type UintGauge = GenericGauge<AtomicU64>;
+
 #[easy_ext::ext(IntGaugeExt)]
 impl IntGauge {
     /// Increment the gauge, and return a guard that will decrement the gauge when dropped.
@@ -39,3 +44,29 @@ impl IntGauge {
         Guard::create(self)
     }
 }
+
+#[easy_ext::ext(UintGaugeExt)]
+impl UintGauge {
+    /// Increment the gauge, and return a guard that will decrement the gauge when dropped.
+    #[must_use]
+    pub fn inc_guard(&self) -> impl Drop + '_ {
+        struct Guard<'a> {
+            gauge: &'a UintGauge,
+        }
+
+        impl<'a> Guard<'a> {
+            fn create(gauge: &'a UintGauge) -> Self {
+                gauge.inc();
+                Self { gauge }
+            }
+        }
+
+        impl Drop for Guard<'_> {
+            fn drop(&mut self) {
+                self.gauge.dec();
+            }
+        }
+
+        Guard::create(self)
+    }
+}
diff --git a/src/common/metrics/src/guarded_metrics.rs b/src/common/metrics/src/guarded_metrics.rs
index 8e953e28ac948..809e205db0ec2 100644
--- a/src/common/metrics/src/guarded_metrics.rs
+++ b/src/common/metrics/src/guarded_metrics.rs
@@ -86,7 +86,7 @@ macro_rules! register_guarded_int_gauge_vec_with_registry {
 #[macro_export]
 macro_rules! register_guarded_uint_gauge_vec_with_registry {
     ($NAME:expr, $HELP:expr, $LABELS_NAMES:expr, $REGISTRY:expr $(,)?) => {{
-        let inner = prometheus::core::GenericGaugeVec::<AtomicU64>::new(
+        let inner = prometheus::core::GenericGaugeVec::<prometheus::core::AtomicU64>::new(
             prometheus::opts!($NAME, $HELP),
             $LABELS_NAMES,
         );
@@ -142,6 +142,8 @@ mod tait {
 }
 pub use tait::*;
 
+use crate::UintGauge;
+
 pub type LabelGuardedHistogramVec<const N: usize> = LabelGuardedMetricVec<VecBuilderOfHistogram, N>;
 pub type LabelGuardedIntCounterVec<const N: usize> =
     LabelGuardedMetricVec<VecBuilderOfCounter<AtomicU64>, N>;
@@ -155,6 +157,7 @@ pub type LabelGuardedGaugeVec<const N: usize> =
 pub type LabelGuardedHistogram<const N: usize> = LabelGuardedMetric<Histogram, N>;
 pub type LabelGuardedIntCounter<const N: usize> = LabelGuardedMetric<IntCounter, N>;
 pub type LabelGuardedIntGauge<const N: usize> = LabelGuardedMetric<IntGauge, N>;
+pub type LabelGuardedUintGauge<const N: usize> = LabelGuardedMetric<UintGauge, N>;
 pub type LabelGuardedGauge<const N: usize> = LabelGuardedMetric<Gauge, N>;
 
 pub type LabelGuardedLocalHistogram<const N: usize> = LabelGuardedMetric<LocalHistogram, N>;
diff --git a/src/common/proc_macro/Cargo.toml b/src/common/proc_macro/Cargo.toml
index 2219dc2efaba4..d44951a4aec56 100644
--- a/src/common/proc_macro/Cargo.toml
+++ b/src/common/proc_macro/Cargo.toml
@@ -10,11 +10,9 @@ repository = { workspace = true }
 [lib]
 proc-macro = true
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
 
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
+
+
 
 [dependencies]
 proc-macro-error = "1.0"
diff --git a/src/common/secret/Cargo.toml b/src/common/secret/Cargo.toml
index 6d501594acbde..ebd042b23a9b0 100644
--- a/src/common/secret/Cargo.toml
+++ b/src/common/secret/Cargo.toml
@@ -7,11 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
 
 [dependencies]
 aes-gcm = "0.10"
diff --git a/src/common/src/util/scan_range.rs b/src/common/src/util/scan_range.rs
index b45db071be3cc..5c56550eed279 100644
--- a/src/common/src/util/scan_range.rs
+++ b/src/common/src/util/scan_range.rs
@@ -12,12 +12,15 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::cmp::Ordering;
 use std::ops::{Bound, RangeBounds};
 
+use itertools::Itertools;
 use paste::paste;
 use risingwave_pb::batch_plan::scan_range::Bound as PbBound;
 use risingwave_pb::batch_plan::ScanRange as PbScanRange;
 
+use super::sort_util::{cmp_rows, OrderType};
 use crate::hash::table_distribution::TableDistribution;
 use crate::hash::VirtualNode;
 use crate::types::{Datum, ScalarImpl};
@@ -98,6 +101,147 @@ impl ScanRange {
             range: full_range(),
         }
     }
+
+    pub fn convert_to_range(&self) -> (Bound<Vec<Datum>>, Bound<Vec<Datum>>) {
+        fn handle_bound(eq_conds: &Vec<Datum>, bound: &Bound<Vec<Datum>>) -> Bound<Vec<Datum>> {
+            match bound {
+                Bound::Included(literal) => {
+                    let mut prefix = eq_conds.clone();
+                    prefix.extend_from_slice(literal);
+                    Bound::Included(prefix)
+                }
+                Bound::Excluded(literal) => {
+                    let mut prefix = eq_conds.clone();
+                    prefix.extend_from_slice(literal);
+                    Bound::Excluded(prefix)
+                }
+                Bound::Unbounded => {
+                    if eq_conds.is_empty() {
+                        Bound::Unbounded
+                    } else {
+                        Bound::Included(eq_conds.clone())
+                    }
+                }
+            }
+        }
+
+        let new_left = handle_bound(&self.eq_conds, &self.range.0);
+        let new_right = handle_bound(&self.eq_conds, &self.range.1);
+        (new_left, new_right)
+    }
+
+    pub fn is_overlap(left: &ScanRange, right: &ScanRange, order_types: &[OrderType]) -> bool {
+        let range_left = left.convert_to_range();
+        let range_right = right.convert_to_range();
+        Self::range_overlap_check(range_left, range_right, order_types)
+    }
+
+    fn range_overlap_check(
+        left: (Bound<Vec<Datum>>, Bound<Vec<Datum>>),
+        right: (Bound<Vec<Datum>>, Bound<Vec<Datum>>),
+        order_types: &[OrderType],
+    ) -> bool {
+        let (left_start, left_end) = &left;
+        let (right_start, right_end) = &right;
+
+        let left_start_vec = match &left_start {
+            Bound::Included(vec) | Bound::Excluded(vec) => vec,
+            _ => &vec![],
+        };
+        let right_start_vec = match &right_start {
+            Bound::Included(vec) | Bound::Excluded(vec) => vec,
+            _ => &vec![],
+        };
+
+        if left_start_vec.is_empty() && right_start_vec.is_empty() {
+            return true;
+        }
+
+        let order_types = if order_types.iter().all(|o| o.is_ascending()) {
+            order_types
+        } else {
+            // reverse order types to ascending
+            &order_types
+                .iter()
+                .cloned()
+                .map(|o| if o.is_descending() { o.reverse() } else { o })
+                .collect_vec()
+        };
+
+        // Unbounded is always less than any other bound
+        if left_start_vec.is_empty() {
+            // pass
+        } else if right_start_vec.is_empty() {
+            return Self::range_overlap_check(right, left, order_types);
+        } else {
+            assert!(!left_start_vec.is_empty());
+            assert!(!right_start_vec.is_empty());
+            let cmp_column_len = left_start_vec.len().min(right_start_vec.len());
+            let cmp_start = cmp_rows(
+                &left_start_vec[0..cmp_column_len],
+                &right_start_vec[0..cmp_column_len],
+                &order_types[0..cmp_column_len],
+            );
+
+            let right_start_before_left_start = cmp_start.is_gt();
+
+            if right_start_before_left_start {
+                return Self::range_overlap_check(right, left, order_types);
+            }
+
+            if cmp_start == Ordering::Equal
+                && let (Bound::Included(_), Bound::Included(_)) = (left_start, right_start)
+            {
+                return true;
+            }
+        }
+
+        let left_end_vec = match &left_end {
+            Bound::Included(vec) | Bound::Excluded(vec) => vec,
+            _ => &vec![],
+        };
+        let right_end_vec = match &right_end {
+            Bound::Included(vec) | Bound::Excluded(vec) => vec,
+            _ => &vec![],
+        };
+
+        if left_end_vec.is_empty() && right_end_vec.is_empty() {
+            return true;
+        }
+
+        if left_end_vec.is_empty() {
+            true
+        } else {
+            // cmp left_end and right_start
+            assert!(!left_end_vec.is_empty());
+            assert!(!right_start_vec.is_empty());
+
+            let cmp_column_len = left_end_vec.len().min(right_start_vec.len());
+            let cmp_end = cmp_rows(
+                &left_end_vec[0..cmp_column_len],
+                &right_start_vec[0..cmp_column_len],
+                &order_types[0..cmp_column_len],
+            );
+
+            match cmp_end {
+                Ordering::Equal => {
+                    if let (Bound::Included(_), Bound::Included(_)) = (left_end, right_start) {
+                        return true;
+                    }
+                }
+
+                Ordering::Greater => {
+                    return true;
+                }
+
+                Ordering::Less => {
+                    return false;
+                }
+            }
+
+            false
+        }
+    }
 }
 
 pub const fn full_range<T>() -> (Bound<T>, Bound<T>) {
@@ -221,4 +365,465 @@ mod tests {
 
         assert_eq!(scan_range.try_compute_vnode(&dist), Some(vnode));
     }
+
+    #[test]
+    fn test_convert_to_range() {
+        {
+            // test empty eq_conds
+            let scan_range = ScanRange {
+                eq_conds: vec![],
+                range: (
+                    Bound::Included(vec![Some(ScalarImpl::from(1))]),
+                    Bound::Included(vec![Some(ScalarImpl::from(2))]),
+                ),
+            };
+
+            let (left, right) = scan_range.convert_to_range();
+            assert_eq!(left, Bound::Included(vec![Some(ScalarImpl::from(1))]));
+            assert_eq!(right, Bound::Included(vec![Some(ScalarImpl::from(2))]));
+        }
+
+        {
+            // test exclude bound with empty eq_conds
+            let scan_range = ScanRange {
+                eq_conds: vec![],
+                range: (
+                    Bound::Excluded(vec![Some(ScalarImpl::from(1))]),
+                    Bound::Excluded(vec![Some(ScalarImpl::from(2))]),
+                ),
+            };
+
+            let (left, right) = scan_range.convert_to_range();
+            assert_eq!(left, Bound::Excluded(vec![Some(ScalarImpl::from(1))]));
+            assert_eq!(right, Bound::Excluded(vec![Some(ScalarImpl::from(2))]));
+        }
+
+        {
+            // test include bound with empty eq_conds
+            let scan_range = ScanRange {
+                eq_conds: vec![],
+                range: (
+                    Bound::Included(vec![Some(ScalarImpl::from(1))]),
+                    Bound::Unbounded,
+                ),
+            };
+
+            let (left, right) = scan_range.convert_to_range();
+            assert_eq!(left, Bound::Included(vec![Some(ScalarImpl::from(1))]));
+            assert_eq!(right, Bound::Unbounded);
+        }
+
+        {
+            // test exclude bound with non-empty eq_conds
+            let scan_range = ScanRange {
+                eq_conds: vec![Some(ScalarImpl::from(1))],
+                range: (
+                    Bound::Excluded(vec![Some(ScalarImpl::from(2))]),
+                    Bound::Excluded(vec![Some(ScalarImpl::from(3))]),
+                ),
+            };
+
+            let (left, right) = scan_range.convert_to_range();
+            assert_eq!(
+                left,
+                Bound::Excluded(vec![Some(ScalarImpl::from(1)), Some(ScalarImpl::from(2))])
+            );
+            assert_eq!(
+                right,
+                Bound::Excluded(vec![Some(ScalarImpl::from(1)), Some(ScalarImpl::from(3))])
+            );
+        }
+
+        {
+            // test include bound with non-empty eq_conds
+            let scan_range = ScanRange {
+                eq_conds: vec![Some(ScalarImpl::from(1))],
+                range: (
+                    Bound::Included(vec![Some(ScalarImpl::from(2))]),
+                    Bound::Included(vec![Some(ScalarImpl::from(3))]),
+                ),
+            };
+
+            let (left, right) = scan_range.convert_to_range();
+            assert_eq!(
+                left,
+                Bound::Included(vec![Some(ScalarImpl::from(1)), Some(ScalarImpl::from(2))])
+            );
+            assert_eq!(
+                right,
+                Bound::Included(vec![Some(ScalarImpl::from(1)), Some(ScalarImpl::from(3))])
+            );
+        }
+
+        {
+            let scan_range = ScanRange {
+                eq_conds: vec![Some(ScalarImpl::from(1))],
+                range: (
+                    Bound::Included(vec![Some(ScalarImpl::from(2))]),
+                    Bound::Unbounded,
+                ),
+            };
+
+            let (left, right) = scan_range.convert_to_range();
+            assert_eq!(
+                left,
+                Bound::Included(vec![Some(ScalarImpl::from(1)), Some(ScalarImpl::from(2))])
+            );
+            assert_eq!(right, Bound::Included(vec![Some(ScalarImpl::from(1))]));
+        }
+    }
+
+    #[test]
+    fn test_range_overlap_check() {
+        let order_types = vec![OrderType::ascending()];
+
+        // (Included, Included) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Included, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Excluded, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Excluded, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Excluded) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Excluded) vs (Included, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Excluded) vs (Excluded, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Excluded) vs (Excluded, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Included) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Included) vs (Included, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Included) vs (Excluded, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Included) vs (Excluded, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Excluded) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Excluded) vs (Included, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Excluded) vs (Excluded, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Excluded, Excluded) vs (Excluded, Excluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3)), Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7)), Some(ScalarImpl::Int32(7))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Included, Included)
+        assert!(!ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(2))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Included, Included)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Excluded, Encluded)
+        assert!(!ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))])
+            ),
+            (
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(3))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        // (Included, Included) vs (Included, Encluded)
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Excluded(vec![Some(ScalarImpl::Int32(7))])
+            ),
+            &order_types
+        ));
+
+        assert!(!ScanRange::range_overlap_check(
+            (
+                Bound::Unbounded,
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))]),
+                Bound::Unbounded,
+            ),
+            &order_types
+        ));
+
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Unbounded,
+                Bound::Included(vec![Some(ScalarImpl::Int32(10))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))]),
+                Bound::Unbounded,
+            ),
+            &order_types
+        ));
+
+        assert!(ScanRange::range_overlap_check(
+            (Bound::Unbounded, Bound::Unbounded,),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))]),
+                Bound::Unbounded,
+            ),
+            &order_types
+        ));
+
+        assert!(ScanRange::range_overlap_check(
+            (Bound::Unbounded, Bound::Unbounded),
+            (Bound::Unbounded, Bound::Unbounded),
+            &order_types
+        ));
+
+        assert!(!ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))])
+            ),
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))]),
+                Bound::Unbounded,
+            ),
+            &order_types
+        ));
+
+        assert!(ScanRange::range_overlap_check(
+            (
+                Bound::Included(vec![Some(ScalarImpl::Int32(1))]),
+                Bound::Included(vec![Some(ScalarImpl::Int32(3))])
+            ),
+            (
+                Bound::Unbounded,
+                Bound::Included(vec![Some(ScalarImpl::Int32(5))]),
+            ),
+            &order_types
+        ));
+    }
 }
diff --git a/src/common/telemetry_event/Cargo.toml b/src/common/telemetry_event/Cargo.toml
index 6007f1cdc3314..ab3826b9504b0 100644
--- a/src/common/telemetry_event/Cargo.toml
+++ b/src/common/telemetry_event/Cargo.toml
@@ -7,18 +7,11 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 jsonbb = { workspace = true }
 prost = { workspace = true }
 reqwest = { version = "0.12.2", features = ["json"] }
 risingwave_pb = { workspace = true }
-thiserror = "1"
 thiserror-ext = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio", features = [
     "rt",
@@ -29,4 +22,3 @@ tokio = { version = "0.2", package = "madsim-tokio", features = [
     "signal",
 ] }
 tracing = "0.1"
-uuid = { version = "1", features = ["v4"] }
diff --git a/src/compute/Cargo.toml b/src/compute/Cargo.toml
index ef66755e3e5be..d6a14aa04cd9e 100644
--- a/src/compute/Cargo.toml
+++ b/src/compute/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 async-trait = "0.1"
@@ -24,7 +18,6 @@ foyer = { workspace = true }
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
 futures-async-stream = { workspace = true }
 http = "1"
-hyper = "1"
 itertools = { workspace = true }
 maplit = "1.0.2"
 pprof = { version = "0.14", features = ["flamegraph"] }
@@ -60,7 +53,6 @@ tokio-stream = { workspace = true }
 tonic = { workspace = true }
 tower = { version = "0.5", features = ["util", "load-shed"] }
 tracing = "0.1"
-uuid = { version = "1.11.0", features = ["v4"] }
 
 [target.'cfg(not(madsim))'.dependencies]
 workspace-hack = { path = "../workspace-hack" }
diff --git a/src/compute/src/rpc/service/monitor_service.rs b/src/compute/src/rpc/service/monitor_service.rs
index 2671fd533262f..cf0b76919fb59 100644
--- a/src/compute/src/rpc/service/monitor_service.rs
+++ b/src/compute/src/rpc/service/monitor_service.rs
@@ -21,16 +21,18 @@ use std::time::Duration;
 use foyer::{HybridCache, TracingOptions};
 use itertools::Itertools;
 use prometheus::core::Collector;
+use prometheus::proto::Metric;
 use risingwave_common::config::{MetricLevel, ServerConfig};
 use risingwave_common_heap_profiling::{AUTO_DUMP_SUFFIX, COLLAPSED_SUFFIX, MANUALLY_DUMP_SUFFIX};
 use risingwave_hummock_sdk::HummockSstableObjectId;
 use risingwave_jni_core::jvm_runtime::dump_jvm_stack_traces;
 use risingwave_pb::monitor_service::monitor_service_server::MonitorService;
 use risingwave_pb::monitor_service::{
-    AnalyzeHeapRequest, AnalyzeHeapResponse, BackPressureInfo, GetBackPressureRequest,
-    GetBackPressureResponse, HeapProfilingRequest, HeapProfilingResponse, ListHeapProfilingRequest,
-    ListHeapProfilingResponse, ProfilingRequest, ProfilingResponse, StackTraceRequest,
-    StackTraceResponse, TieredCacheTracingRequest, TieredCacheTracingResponse,
+    AnalyzeHeapRequest, AnalyzeHeapResponse, BackPressureInfo, FragmentStats,
+    GetBackPressureRequest, GetBackPressureResponse, HeapProfilingRequest, HeapProfilingResponse,
+    ListHeapProfilingRequest, ListHeapProfilingResponse, ProfilingRequest, ProfilingResponse,
+    RelationStats, StackTraceRequest, StackTraceResponse, TieredCacheTracingRequest,
+    TieredCacheTracingResponse,
 };
 use risingwave_rpc_client::error::ToTonicStatus;
 use risingwave_storage::hummock::compactor::await_tree_key::Compaction;
@@ -309,19 +311,77 @@ impl MonitorService for MonitorServiceImpl {
 
         let actor_count: HashMap<_, _> = actor_count
             .iter()
-            .filter_map(|m| {
-                let fragment_id = m
-                    .get_label()
-                    .iter()
-                    .find(|lp| lp.get_name() == "fragment_id")?
-                    .get_value()
-                    .parse::<u32>()
-                    .unwrap();
+            .map(|m| {
+                let fragment_id = get_label(m, "fragment_id").unwrap();
                 let count = m.get_gauge().get_value() as u32;
-                Some((fragment_id, count))
+                (fragment_id, count)
             })
             .collect();
 
+        let mut fragment_stats: HashMap<u32, FragmentStats> = HashMap::new();
+        for (&fragment_id, &actor_count) in &actor_count {
+            fragment_stats.insert(
+                fragment_id,
+                FragmentStats {
+                    actor_count,
+                    current_epoch: 0,
+                },
+            );
+        }
+
+        let actor_current_epoch = metrics
+            .actor_current_epoch
+            .collect()
+            .into_iter()
+            .next()
+            .unwrap()
+            .take_metric();
+        for m in &actor_current_epoch {
+            let fragment_id = get_label(m, "fragment_id").unwrap();
+            let epoch = m.get_gauge().get_value() as u64;
+            if let Some(s) = fragment_stats.get_mut(&fragment_id) {
+                s.current_epoch = if s.current_epoch == 0 {
+                    epoch
+                } else {
+                    u64::min(s.current_epoch, epoch)
+                }
+            } else {
+                warn!(
+                    fragment_id = fragment_id,
+                    "Miss corresponding actor count metrics"
+                );
+            }
+        }
+
+        let mut relation_stats: HashMap<u32, RelationStats> = HashMap::new();
+        let mview_current_epoch = metrics
+            .materialize_current_epoch
+            .collect()
+            .into_iter()
+            .next()
+            .unwrap()
+            .take_metric();
+        for m in &mview_current_epoch {
+            let table_id = get_label(m, "table_id").unwrap();
+            let epoch = m.get_gauge().get_value() as u64;
+            if let Some(s) = relation_stats.get_mut(&table_id) {
+                s.current_epoch = if s.current_epoch == 0 {
+                    epoch
+                } else {
+                    u64::min(s.current_epoch, epoch)
+                };
+                s.actor_count += 1;
+            } else {
+                relation_stats.insert(
+                    table_id,
+                    RelationStats {
+                        actor_count: 1,
+                        current_epoch: epoch,
+                    },
+                );
+            }
+        }
+
         let mut back_pressure_infos: HashMap<_, BackPressureInfo> = HashMap::new();
 
         for label_pairs in actor_output_buffer_blocking_duration_ns {
@@ -360,6 +420,8 @@ impl MonitorService for MonitorServiceImpl {
 
         Ok(Response::new(GetBackPressureResponse {
             back_pressure_infos: back_pressure_infos.into_values().collect(),
+            fragment_stats,
+            relation_stats,
         }))
     }
 
@@ -436,6 +498,16 @@ impl MonitorService for MonitorServiceImpl {
     }
 }
 
+fn get_label(metric: &Metric, label: &str) -> Option<u32> {
+    metric
+        .get_label()
+        .iter()
+        .find(|lp| lp.get_name() == label)?
+        .get_value()
+        .parse::<u32>()
+        .ok()
+}
+
 pub use grpc_middleware::*;
 
 pub mod grpc_middleware {
diff --git a/src/connector/Cargo.toml b/src/connector/Cargo.toml
index 3f12926c9edac..b1ab8b9a8a72c 100644
--- a/src/connector/Cargo.toml
+++ b/src/connector/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 apache-avro = { workspace = true }
@@ -61,12 +55,7 @@ google-cloud-bigquery = { version = "0.13.0", features = ["auth"] }
 google-cloud-gax = "0.19.0"
 google-cloud-googleapis = { version = "0.15", features = ["pubsub", "bigquery"] }
 google-cloud-pubsub = "0.29"
-governor = { version = "0.6", default-features = false, features = [
-  "std",
-  "dashmap",
-  "jitter",
-] }
-http = "0.2"
+governor = { workspace = true }
 iceberg = { workspace = true }
 iceberg-catalog-glue = { workspace = true }
 iceberg-catalog-rest = { workspace = true }
@@ -74,8 +63,6 @@ icelake = { workspace = true }
 indexmap = { version = "2.7.0", features = ["serde"] }
 itertools = { workspace = true }
 jni = { version = "0.21.1", features = ["invocation"] }
-jsonbb = { workspace = true }
-jsonwebtoken = "9.2.0"
 maplit = "1.0.2"
 moka = { version = "0.12.8", features = ["future"] }
 mongodb = { version = "2.8.2", features = ["tokio-runtime"] }
@@ -125,7 +112,6 @@ risingwave_common = { workspace = true }
 risingwave_common_estimate_size = { workspace = true }
 risingwave_connector_codec = { workspace = true }
 risingwave_jni_core = { workspace = true }
-risingwave_object_store = { workspace = true }
 risingwave_pb = { workspace = true }
 risingwave_rpc_client = { workspace = true }
 rumqttc = { version = "0.24.0", features = ["url"] }
diff --git a/src/connector/codec/Cargo.toml b/src/connector/codec/Cargo.toml
index 9038c928dc3da..50b3a53887404 100644
--- a/src/connector/codec/Cargo.toml
+++ b/src/connector/codec/Cargo.toml
@@ -8,12 +8,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 apache-avro = { workspace = true }
diff --git a/src/connector/src/sink/mongodb.rs b/src/connector/src/sink/mongodb.rs
index 244a1bb70db8d..44d116cedf093 100644
--- a/src/connector/src/sink/mongodb.rs
+++ b/src/connector/src/sink/mongodb.rs
@@ -81,6 +81,12 @@ mod send_bulk_write_command_future {
             )))
         })?;
 
+        if let Ok(ok) = result.get_i32("ok")
+            && ok != 1
+        {
+            return Err(SinkError::Mongodb(anyhow!("bulk write write errors")));
+        }
+
         if let Ok(write_errors) = result.get_array("writeErrors") {
             return Err(SinkError::Mongodb(anyhow!(
                 "bulk write respond with write errors: {:?}",
@@ -88,15 +94,10 @@ mod send_bulk_write_command_future {
             )));
         }
 
-        let n = result.get_i32("n").map_err(|err| {
-            SinkError::Mongodb(
-                anyhow!(err).context("can't extract field n from bulk write response"),
-            )
-        })?;
-        if n < 1 {
+        if let Ok(write_concern_error) = result.get_array("writeConcernError") {
             return Err(SinkError::Mongodb(anyhow!(
-                "bulk write respond with an abnormal state, n = {}",
-                n
+                "bulk write respond with write errors: {:?}",
+                write_concern_error,
             )));
         }
 
diff --git a/src/connector/src/sink/postgres.rs b/src/connector/src/sink/postgres.rs
index 02abe81f020e1..f077e9ac42d9d 100644
--- a/src/connector/src/sink/postgres.rs
+++ b/src/connector/src/sink/postgres.rs
@@ -13,48 +13,30 @@
 // limitations under the License.
 
 use std::collections::{BTreeMap, HashSet};
-use std::sync::Arc;
 
-use anyhow::{anyhow, Context};
+use anyhow::anyhow;
 use async_trait::async_trait;
 use itertools::Itertools;
 use risingwave_common::array::{Op, StreamChunk};
-use risingwave_common::bitmap::Bitmap;
+use risingwave_common::bail;
 use risingwave_common::catalog::Schema;
 use risingwave_common::row::{Row, RowExt};
 use serde_derive::Deserialize;
 use serde_with::{serde_as, DisplayFromStr};
 use simd_json::prelude::ArrayTrait;
 use thiserror_ext::AsReport;
-use tokio_postgres::Statement;
+use tokio_postgres::types::Type as PgType;
 
 use super::{
-    SinkError, SinkWriterMetrics, SINK_TYPE_APPEND_ONLY, SINK_TYPE_OPTION, SINK_TYPE_UPSERT,
+    LogSinker, SinkError, SinkLogReader, SINK_TYPE_APPEND_ONLY, SINK_TYPE_OPTION, SINK_TYPE_UPSERT,
 };
 use crate::connector_common::{create_pg_client, PostgresExternalTable, SslMode};
 use crate::parser::scalar_adapter::{validate_pg_type_to_rw_type, ScalarAdapter};
-use crate::sink::writer::{LogSinkerOf, SinkWriter, SinkWriterExt};
+use crate::sink::log_store::{LogStoreReadItem, TruncateOffset};
 use crate::sink::{DummySinkCommitCoordinator, Result, Sink, SinkParam, SinkWriterParam};
 
 pub const POSTGRES_SINK: &str = "postgres";
 
-macro_rules! rw_row_to_pg_values {
-    ($row:expr, $statement:expr) => {
-        $row.iter().enumerate().map(|(i, d)| {
-            d.and_then(|d| {
-                let ty = &$statement.params()[i];
-                match ScalarAdapter::from_scalar(d, ty) {
-                    Ok(scalar) => Some(scalar),
-                    Err(e) => {
-                        tracing::error!(error=%e.as_report(), scalar=?d, "Failed to convert scalar to pg value");
-                        None
-                    }
-                }
-            })
-        })
-    };
-}
-
 #[serde_as]
 #[derive(Clone, Debug, Deserialize)]
 pub struct PostgresConfig {
@@ -143,7 +125,7 @@ impl TryFrom<SinkParam> for PostgresSink {
 
 impl Sink for PostgresSink {
     type Coordinator = DummySinkCommitCoordinator;
-    type LogSinker = LogSinkerOf<PostgresSinkWriter>;
+    type LogSinker = PostgresSinkWriter;
 
     const SINK_NAME: &'static str = POSTGRES_SINK;
 
@@ -239,40 +221,78 @@ impl Sink for PostgresSink {
         Ok(())
     }
 
-    async fn new_log_sinker(&self, writer_param: SinkWriterParam) -> Result<Self::LogSinker> {
-        Ok(PostgresSinkWriter::new(
+    async fn new_log_sinker(&self, _writer_param: SinkWriterParam) -> Result<Self::LogSinker> {
+        PostgresSinkWriter::new(
             self.config.clone(),
             self.schema.clone(),
             self.pk_indices.clone(),
             self.is_append_only,
         )
-        .await?
-        .into_log_sinker(SinkWriterMetrics::new(&writer_param)))
+        .await
     }
 }
 
-struct Buffer {
-    buffer: Vec<StreamChunk>,
-    size: usize,
+struct ParameterBuffer<'a> {
+    /// A set of parameters to be inserted/deleted.
+    /// Each set is a flattened 2d-array.
+    parameters: Vec<Vec<Option<ScalarAdapter>>>,
+    /// the column dimension (fixed).
+    column_length: usize,
+    /// schema types to serialize into `ScalarAdapter`
+    schema_types: &'a [PgType],
+    /// estimated number of parameters that can be sent in a single query.
+    estimated_parameter_size: usize,
+    /// current parameter buffer to be filled.
+    current_parameter_buffer: Vec<Option<ScalarAdapter>>,
 }
 
-impl Buffer {
-    fn new() -> Self {
+impl<'a> ParameterBuffer<'a> {
+    /// The maximum number of parameters that can be sent in a single query.
+    /// See: <https://www.postgresql.org/docs/current/limits.html>
+    /// and <https://github.com/sfackler/rust-postgres/issues/356>
+    const MAX_PARAMETERS: usize = 32768;
+
+    /// `flattened_chunk_size` is the number of datums in a single chunk.
+    fn new(schema_types: &'a [PgType], flattened_chunk_size: usize) -> Self {
+        let estimated_parameter_size = usize::min(Self::MAX_PARAMETERS, flattened_chunk_size);
         Self {
-            buffer: Vec::new(),
-            size: 0,
+            parameters: vec![],
+            column_length: schema_types.len(),
+            schema_types,
+            estimated_parameter_size,
+            current_parameter_buffer: Vec::with_capacity(estimated_parameter_size),
         }
     }
 
-    fn push(&mut self, chunk: StreamChunk) -> usize {
-        self.size += chunk.cardinality();
-        self.buffer.push(chunk);
-        self.size
+    fn add_row(&mut self, row: impl Row) {
+        if self.current_parameter_buffer.len() + self.column_length >= Self::MAX_PARAMETERS {
+            self.new_buffer();
+        }
+        for (i, datum_ref) in row.iter().enumerate() {
+            let pg_datum = datum_ref.map(|s| {
+                let ty = &self.schema_types[i];
+                match ScalarAdapter::from_scalar(s, ty) {
+                    Ok(scalar) => Some(scalar),
+                    Err(e) => {
+                        tracing::error!(error=%e.as_report(), scalar=?s, "Failed to convert scalar to pg value");
+                        None
+                    }
+                }
+            });
+            self.current_parameter_buffer.push(pg_datum.flatten());
+        }
+    }
+
+    fn new_buffer(&mut self) {
+        let filled_buffer = std::mem::replace(
+            &mut self.current_parameter_buffer,
+            Vec::with_capacity(self.estimated_parameter_size),
+        );
+        self.parameters.push(filled_buffer);
     }
 
-    fn drain(&mut self) -> Vec<StreamChunk> {
-        self.size = 0;
-        std::mem::take(&mut self.buffer)
+    fn into_parts(self) -> (Vec<Vec<Option<ScalarAdapter>>>, Vec<Option<ScalarAdapter>>) {
+        (self.parameters, self.current_parameter_buffer)
     }
 }
 
@@ -281,10 +301,8 @@ pub struct PostgresSinkWriter {
     pk_indices: Vec<usize>,
     is_append_only: bool,
     client: tokio_postgres::Client,
-    buffer: Buffer,
-    insert_statement: Statement,
-    delete_statement: Option<Statement>,
-    upsert_statement: Option<Statement>,
+    schema_types: Vec<PgType>,
+    schema: Schema,
 }
 
 impl PostgresSinkWriter {
@@ -338,201 +356,227 @@ impl PostgresSinkWriter {
             schema_types
         };
 
-        let insert_statement = {
-            let insert_sql = create_insert_sql(&schema, &config.table);
-            client
-                .prepare_typed(&insert_sql, &schema_types)
-                .await
-                .context("Failed to prepare insert statement")?
-        };
-
-        let delete_statement = if is_append_only {
-            None
-        } else {
-            let delete_types = pk_indices
-                .iter()
-                .map(|i| schema_types[*i].clone())
-                .collect_vec();
-            let delete_sql = create_delete_sql(&schema, &config.table, &pk_indices);
-            Some(
-                client
-                    .prepare_typed(&delete_sql, &delete_types)
-                    .await
-                    .context("Failed to prepare delete statement")?,
-            )
-        };
-
-        let upsert_statement = if is_append_only {
-            None
-        } else {
-            let upsert_sql = create_upsert_sql(&schema, &config.table, &pk_indices);
-            Some(
-                client
-                    .prepare_typed(&upsert_sql, &schema_types)
-                    .await
-                    .context("Failed to prepare upsert statement")?,
-            )
-        };
-
         let writer = Self {
             config,
             pk_indices,
             is_append_only,
             client,
-            buffer: Buffer::new(),
-            insert_statement,
-            delete_statement,
-            upsert_statement,
+            schema_types,
+            schema,
         };
         Ok(writer)
     }
 
-    async fn flush(&mut self) -> Result<()> {
+    async fn write_batch(&mut self, chunk: StreamChunk) -> Result<()> {
+        // https://www.postgresql.org/docs/current/limits.html
+        // We have a limit of 65,535 parameters in a single query, as restricted by the PostgreSQL protocol.
         if self.is_append_only {
-            for chunk in self.buffer.drain() {
-                for (op, row) in chunk.rows() {
-                    match op {
-                        Op::Insert => {
-                            self.client
-                                .execute_raw(
-                                    &self.insert_statement,
-                                    rw_row_to_pg_values!(row, self.insert_statement),
-                                )
-                                .await?;
-                        }
-                        Op::UpdateInsert | Op::Delete | Op::UpdateDelete => {
-                            debug_assert!(!self.is_append_only);
-                        }
-                    }
-                }
-            }
+            self.write_batch_append_only(chunk).await
         } else {
-            let mut unmatched_update_insert = 0;
-            for chunk in self.buffer.drain() {
-                for (op, row) in chunk.rows() {
-                    match op {
-                        Op::Insert => {
-                            self.client
-                                .execute_raw(
-                                    &self.insert_statement,
-                                    rw_row_to_pg_values!(row, self.insert_statement),
-                                )
-                                .await?;
-                        }
-                        Op::UpdateInsert => {
-                            unmatched_update_insert += 1;
-                            self.client
-                                .execute_raw(
-                                    self.upsert_statement.as_ref().unwrap(),
-                                    rw_row_to_pg_values!(
-                                        row,
-                                        self.upsert_statement.as_ref().unwrap()
-                                    ),
-                                )
-                                .await?;
-                        }
-                        Op::Delete => {
-                            self.client
-                                .execute_raw(
-                                    self.delete_statement.as_ref().unwrap(),
-                                    rw_row_to_pg_values!(
-                                        row.project(&self.pk_indices),
-                                        self.delete_statement.as_ref().unwrap()
-                                    ),
-                                )
-                                .await?;
-                        }
-                        Op::UpdateDelete => {
-                            unmatched_update_insert -= 1;
-                        }
-                    }
+            self.write_batch_non_append_only(chunk).await
+        }
+    }
+
+    async fn write_batch_append_only(&mut self, chunk: StreamChunk) -> Result<()> {
+        let mut transaction = self.client.transaction().await?;
+        // 1d flattened array of parameters to be inserted.
+        let mut parameter_buffer = ParameterBuffer::new(
+            &self.schema_types,
+            chunk.cardinality() * chunk.data_types().len(),
+        );
+        for (op, row) in chunk.rows() {
+            match op {
+                Op::Insert => {
+                    parameter_buffer.add_row(row);
+                }
+                Op::UpdateInsert | Op::Delete | Op::UpdateDelete => {
+                    bail!("append-only sink should not receive update insert, update delete and delete operations")
                 }
             }
-            assert_eq!(unmatched_update_insert, 0);
         }
+        let (parameters, remaining) = parameter_buffer.into_parts();
+        Self::execute_parameter(
+            Op::Insert,
+            &mut transaction,
+            &self.schema,
+            &self.config.table,
+            &self.pk_indices,
+            parameters,
+            remaining,
+        )
+        .await?;
+        transaction.commit().await?;
 
         Ok(())
     }
-}
-
-#[async_trait]
-impl SinkWriter for PostgresSinkWriter {
-    async fn begin_epoch(&mut self, _epoch: u64) -> Result<()> {
-        Ok(())
-    }
 
-    async fn write_batch(&mut self, chunk: StreamChunk) -> Result<()> {
-        let cardinality = self.buffer.push(chunk);
-        if cardinality >= self.config.max_batch_rows {
-            self.flush().await?;
+    async fn write_batch_non_append_only(&mut self, chunk: StreamChunk) -> Result<()> {
+        let mut transaction = self.client.transaction().await?;
+        // 1d flattened array of parameters to be inserted.
+        let mut insert_parameter_buffer = ParameterBuffer::new(
+            &self.schema_types,
+            chunk.cardinality() * chunk.data_types().len(),
+        );
+        let mut delete_parameter_buffer = ParameterBuffer::new(
+            &self.schema_types,
+            chunk.cardinality() * self.pk_indices.len(),
+        );
+        // 1d flattened array of parameters to be deleted.
+        for (op, row) in chunk.rows() {
+            match op {
+                Op::UpdateInsert | Op::UpdateDelete => {
+                    bail!("UpdateInsert and UpdateDelete should have been normalized by the sink executor")
+                }
+                Op::Insert => {
+                    insert_parameter_buffer.add_row(row);
+                }
+                Op::Delete => {
+                    delete_parameter_buffer.add_row(row.project(&self.pk_indices));
+                }
+            }
         }
-        Ok(())
-    }
 
-    async fn barrier(&mut self, is_checkpoint: bool) -> Result<Self::CommitMetadata> {
-        if is_checkpoint {
-            self.flush().await?;
-        }
-        Ok(())
-    }
+        let (delete_parameters, delete_remaining_parameter) = delete_parameter_buffer.into_parts();
+        Self::execute_parameter(
+            Op::Delete,
+            &mut transaction,
+            &self.schema,
+            &self.config.table,
+            &self.pk_indices,
+            delete_parameters,
+            delete_remaining_parameter,
+        )
+        .await?;
+        let (insert_parameters, insert_remaining_parameter) = insert_parameter_buffer.into_parts();
+        Self::execute_parameter(
+            Op::Insert,
+            &mut transaction,
+            &self.schema,
+            &self.config.table,
+            &self.pk_indices,
+            insert_parameters,
+            insert_remaining_parameter,
+        )
+        .await?;
+        transaction.commit().await?;
 
-    async fn abort(&mut self) -> Result<()> {
         Ok(())
     }
 
-    async fn update_vnode_bitmap(&mut self, _vnode_bitmap: Arc<Bitmap>) -> Result<()> {
+    async fn execute_parameter(
+        op: Op,
+        transaction: &mut tokio_postgres::Transaction<'_>,
+        schema: &Schema,
+        table_name: &str,
+        pk_indices: &[usize],
+        parameters: Vec<Vec<Option<ScalarAdapter>>>,
+        remaining_parameter: Vec<Option<ScalarAdapter>>,
+    ) -> Result<()> {
+        let column_length = schema.fields().len();
+        if !parameters.is_empty() {
+            let parameter_length = parameters[0].len();
+            let rows_length = parameter_length / column_length;
+            assert_eq!(
+                parameter_length % column_length,
+                0,
+                "flattened parameters are unaligned"
+            );
+            let statement = match op {
+                Op::Insert => create_insert_sql(schema, table_name, rows_length),
+                Op::Delete => create_delete_sql(schema, table_name, pk_indices, rows_length),
+                _ => unreachable!(),
+            };
+            let statement = transaction.prepare(&statement).await?;
+            for parameter in parameters {
+                transaction.execute_raw(&statement, parameter).await?;
+            }
+        }
+        if !remaining_parameter.is_empty() {
+            let rows_length = remaining_parameter.len() / column_length;
+            assert_eq!(
+                remaining_parameter.len() % column_length,
+                0,
+                "flattened parameters are unaligned"
+            );
+            let statement = match op {
+                Op::Insert => create_insert_sql(schema, table_name, rows_length),
+                Op::Delete => create_delete_sql(schema, table_name, pk_indices, rows_length),
+                _ => unreachable!(),
+            };
+            let statement = transaction.prepare(&statement).await?;
+            transaction
+                .execute_raw(&statement, remaining_parameter)
+                .await?;
+        }
         Ok(())
     }
 }
 
-fn create_insert_sql(schema: &Schema, table_name: &str) -> String {
-    let columns: String = schema
-        .fields()
-        .iter()
-        .map(|field| field.name.clone())
-        .collect_vec()
-        .join(", ");
-    let parameters: String = (0..schema.fields().len())
-        .map(|i| format!("${}", i + 1))
-        .collect_vec()
-        .join(", ");
-    format!("INSERT INTO {table_name} ({columns}) VALUES ({parameters})")
+#[async_trait]
+impl LogSinker for PostgresSinkWriter {
+    async fn consume_log_and_sink(mut self, log_reader: &mut impl SinkLogReader) -> Result<!> {
+        loop {
+            let (epoch, item) = log_reader.next_item().await?;
+            match item {
+                LogStoreReadItem::StreamChunk { chunk, chunk_id } => {
+                    self.write_batch(chunk).await?;
+                    log_reader.truncate(TruncateOffset::Chunk { epoch, chunk_id })?;
+                }
+                LogStoreReadItem::Barrier { .. } => {
+                    log_reader.truncate(TruncateOffset::Barrier { epoch })?;
+                }
+                LogStoreReadItem::UpdateVnodeBitmap(_) => {}
+            }
+        }
+    }
 }
 
-fn create_upsert_sql(schema: &Schema, table_name: &str, pk_indices: &[usize]) -> String {
+fn create_insert_sql(schema: &Schema, table_name: &str, number_of_rows: usize) -> String {
+    let number_of_columns = schema.len();
     let columns: String = schema
         .fields()
         .iter()
         .map(|field| field.name.clone())
         .collect_vec()
         .join(", ");
-    let parameters: String = (0..schema.fields().len())
-        .map(|i| format!("${}", i + 1))
-        .collect_vec()
-        .join(", ");
-    let pk_columns = pk_indices
-        .iter()
-        .map(|i| schema.fields()[*i].name.clone())
-        .collect_vec()
-        .join(", ");
-    let update_parameters: String = (0..schema.fields().len())
-        .filter(|i| !pk_indices.contains(i))
+    let parameters: String = (0..number_of_rows)
         .map(|i| {
-            let column = schema.fields()[i].name.clone();
-            let param = format!("${}", i + 1);
-            format!("{column} = {param}")
+            let row_parameters = (0..number_of_columns)
+                .map(|j| format!("${}", i * number_of_columns + j + 1))
+                .collect_vec()
+                .join(", ");
+            format!("({row_parameters})")
         })
         .collect_vec()
         .join(", ");
-    format!("INSERT INTO {table_name} ({columns}) VALUES ({parameters}) on conflict ({pk_columns}) do update set {update_parameters}")
+    format!("INSERT INTO {table_name} ({columns}) VALUES {parameters}")
 }
 
-fn create_delete_sql(schema: &Schema, table_name: &str, pk_indices: &[usize]) -> String {
-    let parameters: String = pk_indices
-        .iter()
-        .map(|i| format!("{} = ${}", schema.fields()[*i].name, i + 1))
+fn create_delete_sql(
+    schema: &Schema,
+    table_name: &str,
+    pk_indices: &[usize],
+    number_of_rows: usize,
+) -> String {
+    let number_of_pk = pk_indices.len();
+    let parameters: String = (0..number_of_rows)
+        .map(|i| {
+            let row_parameters: String = pk_indices
+                .iter()
+                .enumerate()
+                .map(|(j, pk_index)| {
+                    format!(
+                        "{} = ${}",
+                        schema.fields()[*pk_index].name,
+                        i * number_of_pk + j + 1
+                    )
+                })
+                .collect_vec()
+                .join(" AND ");
+            format!("({row_parameters})")
+        })
         .collect_vec()
-        .join(" AND ");
+        .join(" OR ");
     format!("DELETE FROM {table_name} WHERE {parameters}")
 }
 
@@ -568,10 +612,10 @@ mod tests {
             },
         ]);
         let table_name = "test_table";
-        let sql = create_insert_sql(&schema, table_name);
+        let sql = create_insert_sql(&schema, table_name, 3);
         check(
             sql,
-            expect!["INSERT INTO test_table (a, b) VALUES ($1, $2)"],
+            expect!["INSERT INTO test_table (a, b) VALUES ($1, $2), ($3, $4), ($5, $6)"],
         );
     }
 
@@ -592,31 +636,10 @@ mod tests {
             },
         ]);
         let table_name = "test_table";
-        let sql = create_delete_sql(&schema, table_name, &[1]);
-        check(sql, expect!["DELETE FROM test_table WHERE b = $2"]);
-    }
-
-    #[test]
-    fn test_create_upsert_sql() {
-        let schema = Schema::new(vec![
-            Field {
-                data_type: DataType::Int32,
-                name: "a".to_owned(),
-                sub_fields: vec![],
-                type_name: "".to_owned(),
-            },
-            Field {
-                data_type: DataType::Int32,
-                name: "b".to_owned(),
-                sub_fields: vec![],
-                type_name: "".to_owned(),
-            },
-        ]);
-        let table_name = "test_table";
-        let sql = create_upsert_sql(&schema, table_name, &[1]);
+        let sql = create_delete_sql(&schema, table_name, &[1], 3);
         check(
             sql,
-            expect!["INSERT INTO test_table (a, b) VALUES ($1, $2) on conflict (b) do update set a = $1"],
+            expect!["DELETE FROM test_table WHERE (b = $1) OR (b = $2) OR (b = $3)"],
         );
     }
 }
diff --git a/src/connector/src/sink/remote.rs b/src/connector/src/sink/remote.rs
index 76e65418cdf18..e9b00a1a63e4f 100644
--- a/src/connector/src/sink/remote.rs
+++ b/src/connector/src/sink/remote.rs
@@ -782,7 +782,7 @@ impl EmbeddedConnectorClient {
 
                 match result {
                     Ok(_) => {
-                        tracing::info!("end of jni call {}::{}", class_name, method_name);
+                        tracing::debug!("end of jni call {}::{}", class_name, method_name);
                     }
                     Err(e) => {
                         tracing::error!(error = %e.as_report(), "jni call error");
diff --git a/src/connector/src/source/kafka/enumerator/client.rs b/src/connector/src/source/kafka/enumerator/client.rs
index f77cfac1b9bbe..295450b6907e3 100644
--- a/src/connector/src/source/kafka/enumerator/client.rs
+++ b/src/connector/src/source/kafka/enumerator/client.rs
@@ -20,12 +20,11 @@ use anyhow::{anyhow, Context};
 use async_trait::async_trait;
 use moka::future::Cache as MokaCache;
 use moka::ops::compute::Op;
-use prometheus::core::{AtomicI64, GenericGauge};
 use rdkafka::consumer::{BaseConsumer, Consumer};
 use rdkafka::error::KafkaResult;
 use rdkafka::{ClientConfig, Offset, TopicPartitionList};
 use risingwave_common::bail;
-use risingwave_common::metrics::LabelGuardedMetric;
+use risingwave_common::metrics::LabelGuardedIntGauge;
 
 use crate::error::{ConnectorError, ConnectorResult};
 use crate::source::base::SplitEnumerator;
@@ -60,7 +59,7 @@ pub struct KafkaSplitEnumerator {
     stop_offset: KafkaEnumeratorOffset,
 
     sync_call_timeout: Duration,
-    high_watermark_metrics: HashMap<i32, LabelGuardedMetric<GenericGauge<AtomicI64>, 2>>,
+    high_watermark_metrics: HashMap<i32, LabelGuardedIntGauge<2>>,
 }
 
 impl KafkaSplitEnumerator {}
diff --git a/src/connector/src/source/kafka/source/reader.rs b/src/connector/src/source/kafka/source/reader.rs
index 4c8a78adbfec4..a91407e61e767 100644
--- a/src/connector/src/source/kafka/source/reader.rs
+++ b/src/connector/src/source/kafka/source/reader.rs
@@ -21,12 +21,11 @@ use anyhow::Context;
 use async_trait::async_trait;
 use futures::StreamExt;
 use futures_async_stream::try_stream;
-use prometheus::core::{AtomicI64, GenericGauge};
 use rdkafka::config::RDKafkaLogLevel;
 use rdkafka::consumer::{Consumer, StreamConsumer};
 use rdkafka::error::KafkaError;
 use rdkafka::{ClientConfig, Message, Offset, TopicPartitionList};
-use risingwave_common::metrics::LabelGuardedMetric;
+use risingwave_common::metrics::LabelGuardedIntGauge;
 use risingwave_pb::plan_common::additional_column::ColumnType as AdditionalColumnType;
 
 use crate::error::ConnectorResult as Result;
@@ -247,10 +246,8 @@ impl KafkaSplitReader {
             )
         });
 
-        let mut latest_message_id_metrics: HashMap<
-            String,
-            LabelGuardedMetric<GenericGauge<AtomicI64>, 3>,
-        > = HashMap::new();
+        let mut latest_message_id_metrics: HashMap<String, LabelGuardedIntGauge<3>> =
+            HashMap::new();
 
         #[for_await]
         'for_outer_loop: for msgs in self.consumer.stream().ready_chunks(max_chunk_size) {
diff --git a/src/connector/src/source/kafka/stats.rs b/src/connector/src/source/kafka/stats.rs
index 7a36c4d1fffea..9e8285a8cc486 100644
--- a/src/connector/src/source/kafka/stats.rs
+++ b/src/connector/src/source/kafka/stats.rs
@@ -12,7 +12,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use prometheus::core::AtomicU64;
 use prometheus::Registry;
 use rdkafka::statistics::{Broker, ConsumerGroup, Partition, Topic, Window};
 use rdkafka::Statistics;
diff --git a/src/connector/with_options/Cargo.toml b/src/connector/with_options/Cargo.toml
index 59dd4291e12b7..b8d1ecf510379 100644
--- a/src/connector/with_options/Cargo.toml
+++ b/src/connector/with_options/Cargo.toml
@@ -9,7 +9,6 @@ repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-proc-macro2 = "1"
 quote = "1"
 syn = "2"
 
@@ -18,11 +17,5 @@ proc-macro = true
 
 [dev-dependencies]
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [lints]
 workspace = true
diff --git a/src/ctl/Cargo.toml b/src/ctl/Cargo.toml
index 7b0fc940de946..cdb8f5377c75e 100644
--- a/src/ctl/Cargo.toml
+++ b/src/ctl/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 bytes = "1"
@@ -24,7 +18,6 @@ futures = { version = "0.3", default-features = false, features = ["alloc"] }
 hex = "0.4"
 inquire = "0.7.0"
 itertools = { workspace = true }
-memcomparable = "0.2"
 prost = { workspace = true }
 regex = "1.10.0"
 risingwave_common = { workspace = true }
@@ -33,7 +26,6 @@ risingwave_frontend = { workspace = true }
 risingwave_hummock_sdk = { workspace = true }
 risingwave_meta = { workspace = true }
 risingwave_meta_model = { workspace = true }
-risingwave_meta_model_migration = { workspace = true }
 risingwave_object_store = { workspace = true }
 risingwave_pb = { workspace = true }
 risingwave_rpc_client = { workspace = true }
diff --git a/src/dml/Cargo.toml b/src/dml/Cargo.toml
index df0ce4ebc19d6..1ad438db96259 100644
--- a/src/dml/Cargo.toml
+++ b/src/dml/Cargo.toml
@@ -7,24 +7,13 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
 futures-async-stream = { workspace = true }
 itertools = { workspace = true }
 parking_lot = { workspace = true }
-rand = { workspace = true }
 risingwave_common = { workspace = true }
-risingwave_connector = { workspace = true }
-risingwave_pb = { workspace = true }
-rw_futures_util = { workspace = true }
 thiserror = "1"
-thiserror-ext = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio", features = [
     "rt",
     "rt-multi-thread",
@@ -41,7 +30,6 @@ workspace-hack = { path = "../workspace-hack" }
 
 [dev-dependencies]
 assert_matches = "1"
-criterion = { workspace = true, features = ["async_tokio"] }
 paste = "1"
 tempfile = "3"
 
diff --git a/src/error/Cargo.toml b/src/error/Cargo.toml
index 4a99711db6c41..8407197a88769 100644
--- a/src/error/Cargo.toml
+++ b/src/error/Cargo.toml
@@ -10,7 +10,6 @@ repository = { workspace = true }
 [dependencies]
 anyhow = "1"
 bincode = "1"
-bytes = "1"
 easy-ext = "1"
 serde = "1"
 serde-error = "0.1"
diff --git a/src/expr/core/Cargo.toml b/src/expr/core/Cargo.toml
index 9851186260e5a..16db4ea8f2691 100644
--- a/src/expr/core/Cargo.toml
+++ b/src/expr/core/Cargo.toml
@@ -9,18 +9,11 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack", "ctor"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack", "ctor"]
-
 [dependencies]
 anyhow = "1"
 async-trait = "0.1"
 auto_impl = "1"
 await-tree = { workspace = true }
-cfg-or-panic = "0.2"
 chrono = { version = "0.4", default-features = false, features = [
     "clock",
     "std",
diff --git a/src/expr/impl/Cargo.toml b/src/expr/impl/Cargo.toml
index ec6fa8afb9e99..30bd6f61d6b95 100644
--- a/src/expr/impl/Cargo.toml
+++ b/src/expr/impl/Cargo.toml
@@ -10,13 +10,10 @@ repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [package.metadata.cargo-machete]
-ignored = ["workspace-hack", "ctor"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack", "ctor"]
+ignored = ["chrono-tz", "futures-async-stream"]
 
 [features]
-external-udf = ["arrow-udf-flight", "arrow-flight", "tonic"]
+external-udf = ["arrow-udf-flight", "arrow-flight", "tonic", "ginepro"]
 js-udf = ["arrow-udf-js"]
 python-udf = ["arrow-udf-python"]
 wasm-udf = ["arrow-udf-wasm", "zstd"]
@@ -42,7 +39,7 @@ educe = "0.6"
 fancy-regex = "0.14"
 futures-async-stream = { workspace = true }
 futures-util = "0.3"
-ginepro = "0.8"
+ginepro = { version = "0.8", optional = true }
 hex = "0.4"
 hmac = "0.12"
 iceberg = { workspace = true }
diff --git a/src/frontend/Cargo.toml b/src/frontend/Cargo.toml
index 3041a4320dd42..74fafdd6d7d97 100644
--- a/src/frontend/Cargo.toml
+++ b/src/frontend/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 arc-swap = "1"
@@ -71,7 +65,6 @@ risingwave_dml = { workspace = true }
 risingwave_expr = { workspace = true }
 risingwave_frontend_macro = { path = "macro" }
 risingwave_hummock_sdk = { workspace = true }
-risingwave_object_store = { workspace = true }
 risingwave_pb = { workspace = true }
 risingwave_rpc_client = { workspace = true }
 risingwave_sqlparser = { workspace = true }
@@ -108,7 +101,6 @@ tower-http = { version = "0.6", features = [
 tracing = "0.1"
 url = "2.5.0"
 uuid = "1"
-zstd = { version = "0.13", default-features = false }
 
 [target.'cfg(not(madsim))'.dependencies]
 workspace-hack = { path = "../workspace-hack" }
diff --git a/src/frontend/planner_test/Cargo.toml b/src/frontend/planner_test/Cargo.toml
index 49ad1876afff0..ee8d4af18655d 100644
--- a/src/frontend/planner_test/Cargo.toml
+++ b/src/frontend/planner_test/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 expect-test = "1"
diff --git a/src/frontend/planner_test/tests/testdata/input/range_scan.yaml b/src/frontend/planner_test/tests/testdata/input/range_scan.yaml
index 52863fa5aaf77..374da681d0987 100644
--- a/src/frontend/planner_test/tests/testdata/input/range_scan.yaml
+++ b/src/frontend/planner_test/tests/testdata/input/range_scan.yaml
@@ -6,182 +6,184 @@
             date INTEGER);
     CREATE MATERIALIZED VIEW orders_count_by_user AS
       SELECT user_id, date, count(*) AS orders_count FROM orders GROUP BY user_id, date;
+    CREATE MATERIALIZED VIEW orders_count_by_user_desc AS
+      SELECT user_id, date, count(*) AS orders_count FROM orders GROUP BY user_id, date ORDER BY user_id DESC;
   expected_outputs: []
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id < 43
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 + 1
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1/0
   expected_outputs:
-  - batch_error
+    - batch_error
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 2147483647 + 1
   expected_outputs:
-  - batch_error
+    - batch_error
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 'a'
   expected_outputs:
-  - batch_error
+    - batch_error
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id > 'a'
   expected_outputs:
-  - batch_error
+    - batch_error
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = '42'
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = NULL
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id IS NULL
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id > NULL
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date = 1111
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id > 42 AND date = 1111 AND 2>1
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE date > 1111 AND user_id = 42 AND 5<6 AND date <= 6666
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in (42, 43)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in (42+1, 44-1)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: If the IN list has a larger type than the column, the InputRef is casted. Currently this case is not converted to scan range yet.
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in (42.0, 43.0)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in ('42', '43')
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in ('42', '43.0')
   expected_outputs:
-  - batch_error
+    - batch_error
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in (2147483648, 2147483649) AND date = 6666
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (2222, 3333)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: test duplicate value in in-list
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (2222, 2222)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: test NULL in in-list
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (2222, NULL)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: test NULL in in-list
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (NULL)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: test multiple in-list
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (2222, 3333) AND date in (4444, 3333)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: test eq & in-list
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (2222, 3333) AND date = 3333
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: test eq & in-list
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 42 AND date in (2222, 3333) AND date = 4444
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id in (2147483648, 2147483649) AND date in (2222, 3333)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - id: create_table_and_mv_ordered
   sql: |
     CREATE TABLE orders (
@@ -194,93 +196,93 @@
       ORDER BY orders_count;
   expected_outputs: []
 - before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE user_id = 42
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE user_id > 42 AND orders_count = 10
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count = 10
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge mutiple upper bound
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count < 10 and orders_count < 30
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge include and exclude upper bound of same value
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count < 10 and orders_count <= 10
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge mutiple lower bound
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count > 10 and orders_count > 30
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge include and exclude lower bound of same value
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count > 10 and orders_count >= 10
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: invalid range
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count > 10 and orders_count < 5
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge cmp and eq condition
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count > 20 and orders_count < 30 and orders_count = 25
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: invalid range of merging cmp and eq condition
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count > 20 and orders_count < 30 and orders_count = 35
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge cmp and const-in condition
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count in (10,20,30,40) and orders_count <30
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: invalid range of merging cmp and const-in condition
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count in (10,20,30,40) and orders_count > 50
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: merge null and cmp condition
   before:
-  - create_table_and_mv_ordered
+    - create_table_and_mv_ordered
   sql: |
     SELECT * FROM orders_count_by_user_ordered WHERE orders_count is null and orders_count < 30
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - id: create_small
   sql: |
     CREATE TABLE t(x smallint);
@@ -289,142 +291,241 @@
   expected_outputs: []
 - name: When the constant with larger type is out of the range of the column's type, we can convert it as false condition.
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x = 60000;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When the constant with larger type is out of the upper bound of the column's type, we can convert < as true condition.
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x < 60000;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When the constant with larger type is out of the upper bound of the column's type, we can convert > as false condition.
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x > 60000;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When the constant with larger type is out of the lower bound of the column's type, we can convert < as false condition.
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x < -60000;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When the constant with larger type is out of the lower bound of the column's type, we can convert > as true condition.
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x > -60000;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When the constant with larger type is in range of the column's type, we can convert it.
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x < 3::bigint and x > 1::bigint;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: Can't push down the in-compatitble numeric type
   before:
-  - create_small
+    - create_small
   sql: |
     SELECT * FROM mv WHERE x = 3.4;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1 or user_id = 2;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE (user_id = 1) or (user_id = 2 and date = 2222);
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE (user_id = 1) or (user_id = 2 and date in (1111, 2222));
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When one arm of or clause contains other conditions, we can't convert it to scan range yet.
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE (user_id = 1) or (user_id = 2 and date in (1111, 2222)) or (user_id != 3);
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - name: When any arm of or clause is not equal type, we can't convert it to scan range yet.
   before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id > 1 or user_id < 10
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1 or user_id is null
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1 and user_id is null
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1 or (user_id is null and date = 1111)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1 or (user_id = 2 and date is null)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id = 1 or (user_id is null and date is null)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - before:
-  - create_table_and_mv
+    - create_table_and_mv
   sql: |
     SELECT * FROM orders_count_by_user WHERE user_id is null or (user_id is null and date is null)
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - sql: |
     create table sbtest1(id INT, k INT, c VARCHAR, pad VARCHAR);
     create index k1 on sbtest1(k);
     select count(k) from sbtest1 where k between 0 and 5;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - sql: |
     create table sbtest1(id INT, k INT, c VARCHAR, pad VARCHAR);
     create index k1 on sbtest1(k);
     select count(k) from sbtest1 where k between 0 and 500;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - sql: |
     create table sbtest1(id INT, k INT, c VARCHAR, pad VARCHAR, primary key(id));
     create index k1 on sbtest1(k);
     select count(k) from sbtest1 where id between 0 and 5;
   expected_outputs:
-  - batch_plan
+    - batch_plan
 - sql: |
     create table t (k int primary key, v int);
     select v from t where k = 2147483648; -- out of range of int32
   expected_outputs:
-  - logical_plan
-  - batch_plan
+    - logical_plan
+    - batch_plan
+- name: When OR clauses contain non-overlapping conditions,, we can pushdown serveral scan_range.
+  before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id > 20);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id > 20);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id >= 10);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id >= 10);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id > 20) or (user_id = 15);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id > 20) or (user_id = 15);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id < 30) or (user_id >= 30 and user_id < 40);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id < 30) or (user_id >= 30 and user_id < 40);
+  expected_outputs:
+    - batch_plan
+- name: When OR clauses contain overlapping conditions, we can merge serveral scan_range and pushdown.
+  before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id <= 30) or (user_id >= 30 and user_id < 40) or (user_id = 15);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id <= 30) or (user_id >= 30 and user_id < 40) or (user_id = 15);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id > 30) or (user_id > 5 and user_id < 15);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id > 30) or (user_id > 5 and user_id < 15);
+  expected_outputs:
+    - batch_plan
+- name: When OR clauses contain overlapping conditions, we cannot push down if it results in a full table scan.
+  before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 20) or (user_id > 10);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 20) or (user_id > 10);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 20) or (user_id != 10);
+  expected_outputs:
+    - batch_plan
+- before:
+    - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 20) or (user_id != 10);
+  expected_outputs:
+    - batch_plan
diff --git a/src/frontend/planner_test/tests/testdata/input/row_filter.yaml b/src/frontend/planner_test/tests/testdata/input/row_filter.yaml
index 2b1966f4316fa..7293cadf7512b 100644
--- a/src/frontend/planner_test/tests/testdata/input/row_filter.yaml
+++ b/src/frontend/planner_test/tests/testdata/input/row_filter.yaml
@@ -22,4 +22,10 @@
     create table t(v1 int, v2 int, v3 int, primary key(v1,v2,v3));
     select * from t where (v1,v2,v1) > (1,2,3);
   expected_outputs:
+  - batch_plan
+- sql: |
+    create table t1(v1 int, v2 int, v3 int);
+    create materialized view mv1 as select * from t1 order by v1 asc, v2 asc, v3 desc;
+    select * from mv1 where (v1,v2,v3) > (1,3,1);
+  expected_outputs:
   - batch_plan
\ No newline at end of file
diff --git a/src/frontend/planner_test/tests/testdata/output/range_scan.yaml b/src/frontend/planner_test/tests/testdata/output/range_scan.yaml
index ed8b4b863fe64..e175558038c81 100644
--- a/src/frontend/planner_test/tests/testdata/output/range_scan.yaml
+++ b/src/frontend/planner_test/tests/testdata/output/range_scan.yaml
@@ -7,6 +7,8 @@
             date INTEGER);
     CREATE MATERIALIZED VIEW orders_count_by_user AS
       SELECT user_id, date, count(*) AS orders_count FROM orders GROUP BY user_id, date;
+    CREATE MATERIALIZED VIEW orders_count_by_user_desc AS
+      SELECT user_id, date, count(*) AS orders_count FROM orders GROUP BY user_id, date ORDER BY user_id DESC;
 - before:
   - create_table_and_mv
   sql: |
@@ -503,3 +505,134 @@
     └─LogicalFilter { predicate: (t.k = 2147483648:Int64) }
       └─LogicalScan { table: t, columns: [t.k, t.v, t._rw_timestamp] }
   batch_plan: 'BatchValues { rows: [] }'
+- name: When OR clauses contain non-overlapping conditions,, we can pushdown serveral scan_range.
+  before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id > 20);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user.user_id < 10:Int32) OR (orders_count_by_user.user_id > 20:Int32)) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], scan_ranges: [orders_count_by_user.user_id < Int64(10), orders_count_by_user.user_id > Int64(20)], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id > 20);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user_desc.user_id < 10:Int32) OR (orders_count_by_user_desc.user_id > 20:Int32)) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], scan_ranges: [orders_count_by_user_desc.user_id < Int64(10), orders_count_by_user_desc.user_id > Int64(20)], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id >= 10);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user.user_id < 10:Int32) OR (orders_count_by_user.user_id >= 10:Int32)) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], scan_ranges: [orders_count_by_user.user_id < Int64(10), orders_count_by_user.user_id >= Int64(10)], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id >= 10);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user_desc.user_id < 10:Int32) OR (orders_count_by_user_desc.user_id >= 10:Int32)) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], scan_ranges: [orders_count_by_user_desc.user_id < Int64(10), orders_count_by_user_desc.user_id >= Int64(10)], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id > 20) or (user_id = 15);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: (((orders_count_by_user.user_id < 10:Int32) OR (orders_count_by_user.user_id > 20:Int32)) OR (orders_count_by_user.user_id = 15:Int32)) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], scan_ranges: [orders_count_by_user.user_id < Int64(10), orders_count_by_user.user_id = Int64(15), orders_count_by_user.user_id > Int64(20)], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id > 20) or (user_id = 15);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: (((orders_count_by_user_desc.user_id < 10:Int32) OR (orders_count_by_user_desc.user_id > 20:Int32)) OR (orders_count_by_user_desc.user_id = 15:Int32)) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], scan_ranges: [orders_count_by_user_desc.user_id < Int64(10), orders_count_by_user_desc.user_id = Int64(15), orders_count_by_user_desc.user_id > Int64(20)], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id < 30) or (user_id >= 30 and user_id < 40);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((((orders_count_by_user.user_id < 10:Int32) AND (orders_count_by_user.user_id > 1:Int32)) OR ((orders_count_by_user.user_id > 20:Int32) AND (orders_count_by_user.user_id < 30:Int32))) OR ((orders_count_by_user.user_id >= 30:Int32) AND (orders_count_by_user.user_id < 40:Int32))) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], scan_ranges: [orders_count_by_user.user_id > Int64(1) AND orders_count_by_user.user_id < Int64(10), orders_count_by_user.user_id > Int64(20) AND orders_count_by_user.user_id < Int64(30), orders_count_by_user.user_id >= Int64(30) AND orders_count_by_user.user_id < Int64(40)], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id < 30) or (user_id >= 30 and user_id < 40);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((((orders_count_by_user_desc.user_id < 10:Int32) AND (orders_count_by_user_desc.user_id > 1:Int32)) OR ((orders_count_by_user_desc.user_id > 20:Int32) AND (orders_count_by_user_desc.user_id < 30:Int32))) OR ((orders_count_by_user_desc.user_id >= 30:Int32) AND (orders_count_by_user_desc.user_id < 40:Int32))) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], scan_ranges: [orders_count_by_user_desc.user_id > Int64(1) AND orders_count_by_user_desc.user_id < Int64(10), orders_count_by_user_desc.user_id > Int64(20) AND orders_count_by_user_desc.user_id < Int64(30), orders_count_by_user_desc.user_id >= Int64(30) AND orders_count_by_user_desc.user_id < Int64(40)], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- name: When OR clauses contain overlapping conditions, we can merge serveral scan_range and pushdown.
+  before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id <= 30) or (user_id >= 30 and user_id < 40) or (user_id = 15);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((((orders_count_by_user.user_id < 10:Int32) AND (orders_count_by_user.user_id > 1:Int32)) OR ((orders_count_by_user.user_id > 20:Int32) AND (orders_count_by_user.user_id <= 30:Int32))) OR (((orders_count_by_user.user_id >= 30:Int32) AND (orders_count_by_user.user_id < 40:Int32)) OR (orders_count_by_user.user_id = 15:Int32))) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], scan_ranges: [orders_count_by_user.user_id > Int64(1) AND orders_count_by_user.user_id < Int64(10), orders_count_by_user.user_id = Int64(15), orders_count_by_user.user_id >= Int64(20) AND orders_count_by_user.user_id <= Int64(40)], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10 and user_id > 1) or (user_id > 20 and user_id <= 30) or (user_id >= 30 and user_id < 40) or (user_id = 15);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((((orders_count_by_user_desc.user_id < 10:Int32) AND (orders_count_by_user_desc.user_id > 1:Int32)) OR ((orders_count_by_user_desc.user_id > 20:Int32) AND (orders_count_by_user_desc.user_id <= 30:Int32))) OR (((orders_count_by_user_desc.user_id >= 30:Int32) AND (orders_count_by_user_desc.user_id < 40:Int32)) OR (orders_count_by_user_desc.user_id = 15:Int32))) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], scan_ranges: [orders_count_by_user_desc.user_id > Int64(1) AND orders_count_by_user_desc.user_id < Int64(10), orders_count_by_user_desc.user_id = Int64(15), orders_count_by_user_desc.user_id >= Int64(20) AND orders_count_by_user_desc.user_id <= Int64(40)], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 10) or (user_id > 30) or (user_id > 5 and user_id < 15);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: (((orders_count_by_user.user_id < 10:Int32) OR (orders_count_by_user.user_id > 30:Int32)) OR ((orders_count_by_user.user_id > 5:Int32) AND (orders_count_by_user.user_id < 15:Int32))) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], scan_ranges: [orders_count_by_user.user_id <= Int64(15), orders_count_by_user.user_id > Int64(30)], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 10) or (user_id > 30) or (user_id > 5 and user_id < 15);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: (((orders_count_by_user_desc.user_id < 10:Int32) OR (orders_count_by_user_desc.user_id > 30:Int32)) OR ((orders_count_by_user_desc.user_id > 5:Int32) AND (orders_count_by_user_desc.user_id < 15:Int32))) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], scan_ranges: [orders_count_by_user_desc.user_id <= Int64(15), orders_count_by_user_desc.user_id > Int64(30)], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- name: When OR clauses contain overlapping conditions, we cannot push down if it results in a full table scan.
+  before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 20) or (user_id > 10);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user.user_id < 20:Int32) OR (orders_count_by_user.user_id > 10:Int32)) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 20) or (user_id > 10);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user_desc.user_id < 20:Int32) OR (orders_count_by_user_desc.user_id > 10:Int32)) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user WHERE (user_id < 20) or (user_id != 10);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user.user_id < 20:Int32) OR (orders_count_by_user.user_id <> 10:Int32)) }
+      └─BatchScan { table: orders_count_by_user, columns: [orders_count_by_user.user_id, orders_count_by_user.date, orders_count_by_user.orders_count], distribution: UpstreamHashShard(orders_count_by_user.user_id, orders_count_by_user.date) }
+- before:
+  - create_table_and_mv
+  sql: |
+    SELECT * FROM orders_count_by_user_desc WHERE (user_id < 20) or (user_id != 10);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: ((orders_count_by_user_desc.user_id < 20:Int32) OR (orders_count_by_user_desc.user_id <> 10:Int32)) }
+      └─BatchScan { table: orders_count_by_user_desc, columns: [orders_count_by_user_desc.user_id, orders_count_by_user_desc.date, orders_count_by_user_desc.orders_count], distribution: UpstreamHashShard(orders_count_by_user_desc.user_id, orders_count_by_user_desc.date) }
diff --git a/src/frontend/planner_test/tests/testdata/output/row_filter.yaml b/src/frontend/planner_test/tests/testdata/output/row_filter.yaml
index 1ef80b8e025bd..5de2e79815476 100644
--- a/src/frontend/planner_test/tests/testdata/output/row_filter.yaml
+++ b/src/frontend/planner_test/tests/testdata/output/row_filter.yaml
@@ -17,7 +17,7 @@
   batch_plan: |-
     BatchExchange { order: [], dist: Single }
     └─BatchFilter { predicate: (Row(t.v1, t.v3) > '(2,3)':Struct(StructType { field_names: [], field_types: [Int32, Int32] })) }
-      └─BatchScan { table: t, columns: [t.v1, t.v2, t.v3], scan_ranges: [t.v1 > Int32(2)], distribution: UpstreamHashShard(t.v1, t.v2, t.v3) }
+      └─BatchScan { table: t, columns: [t.v1, t.v2, t.v3], scan_ranges: [t.v1 >= Int32(2)], distribution: UpstreamHashShard(t.v1, t.v2, t.v3) }
 - sql: |
     create table t(v1 int, v2 int, v3 int, primary key(v1,v2,v3));
     select * from t where (v3,v2,v1) > (1,2,3);
@@ -31,4 +31,12 @@
   batch_plan: |-
     BatchExchange { order: [], dist: Single }
     └─BatchFilter { predicate: (Row(t.v1, t.v2, t.v1) > '(1,2,3)':Struct(StructType { field_names: [], field_types: [Int32, Int32, Int32] })) }
-      └─BatchScan { table: t, columns: [t.v1, t.v2, t.v3], scan_ranges: [(t.v1, t.v2) > (Int32(1), Int32(2))], distribution: UpstreamHashShard(t.v1, t.v2, t.v3) }
+      └─BatchScan { table: t, columns: [t.v1, t.v2, t.v3], scan_ranges: [(t.v1, t.v2) >= (Int32(1), Int32(2))], distribution: UpstreamHashShard(t.v1, t.v2, t.v3) }
+- sql: |
+    create table t1(v1 int, v2 int, v3 int);
+    create materialized view mv1 as select * from t1 order by v1 asc, v2 asc, v3 desc;
+    select * from mv1 where (v1,v2,v3) > (1,3,1);
+  batch_plan: |-
+    BatchExchange { order: [], dist: Single }
+    └─BatchFilter { predicate: (Row(mv1.v1, mv1.v2, mv1.v3) > '(1,3,1)':Struct(StructType { field_names: [], field_types: [Int32, Int32, Int32] })) }
+      └─BatchScan { table: mv1, columns: [mv1.v1, mv1.v2, mv1.v3], scan_ranges: [(mv1.v1, mv1.v2) >= (Int32(1), Int32(3))], distribution: SomeShard }
diff --git a/src/frontend/src/catalog/system_catalog/information_schema/key_column_usage.rs b/src/frontend/src/catalog/system_catalog/information_schema/key_column_usage.rs
new file mode 100644
index 0000000000000..c9343b4edcba1
--- /dev/null
+++ b/src/frontend/src/catalog/system_catalog/information_schema/key_column_usage.rs
@@ -0,0 +1,59 @@
+// Copyright 2024 RisingWave Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use risingwave_common::types::Fields;
+use risingwave_frontend_macro::system_catalog;
+
+/// The view `key_column_usage` contains all constraints belonging to tables that the current user owns or has some privilege other than SELECT on.
+/// Ref: [`https://www.postgresql.org/docs/current/infoschema-key-column-usage.html`]
+/// Limitation:
+/// This view assume the constraint schema is the same as the table schema, since `pg_catalog`.`pg_constraint` only support primary key.
+#[system_catalog(
+    view,
+    "information_schema.key_column_usage",
+    "WITH key_column_usage_without_name AS (
+        SELECT CURRENT_DATABASE() AS constraint_catalog,
+            pg_namespace.nspname AS constraint_schema,
+            pg_constraint.conname AS constraint_name,
+            CURRENT_DATABASE() AS table_catalog,
+            pg_namespace.nspname AS table_schema,
+            pg_class.relname AS table_name,
+            unnest(conkey) as col_id,
+            conrelid as table_id
+        FROM pg_catalog.pg_constraint
+        JOIN pg_catalog.pg_class ON pg_constraint.conrelid = pg_class.oid
+        JOIN rw_catalog.rw_relations ON rw_relations.id = pg_class.oid
+        JOIN pg_catalog.pg_namespace ON pg_class.relnamespace = pg_namespace.oid
+        WHERE rw_relations.relation_type != 'table' or (rw_relations.relation_type = 'table' and has_table_privilege(pg_constraint.conrelid, 'INSERT, UPDATE, DELETE'))
+        ORDER BY constraint_catalog, constraint_schema, constraint_name
+    )
+    SELECT constraint_catalog, constraint_schema, constraint_name, table_catalog, table_schema, table_name,
+           name as column_name, rw_columns.position as ordinal_position, NULL::int as position_in_unique_constraint
+    FROM key_column_usage_without_name
+    JOIN rw_catalog.rw_columns ON
+        rw_columns.position = key_column_usage_without_name.col_id AND
+        rw_columns.relation_id = key_column_usage_without_name.table_id"
+)]
+#[derive(Fields)]
+struct KeyColumnUsage {
+    constraint_catalog: String,
+    constraint_schema: String,
+    constraint_name: String,
+    table_catalog: String,
+    table_schema: String,
+    table_name: String,
+    column_name: String,
+    ordinal_position: i32,
+    position_in_unique_constraint: i32,
+}
diff --git a/src/frontend/src/catalog/system_catalog/information_schema/mod.rs b/src/frontend/src/catalog/system_catalog/information_schema/mod.rs
index 6a81d0c109ec0..c3ca755cce489 100644
--- a/src/frontend/src/catalog/system_catalog/information_schema/mod.rs
+++ b/src/frontend/src/catalog/system_catalog/information_schema/mod.rs
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 mod columns;
+mod key_column_usage;
 mod schemata;
 mod table_constraints;
 mod tables;
diff --git a/src/frontend/src/catalog/system_catalog/information_schema/table_constraints.rs b/src/frontend/src/catalog/system_catalog/information_schema/table_constraints.rs
index 24c3970a02a82..3f9a5059a51bc 100644
--- a/src/frontend/src/catalog/system_catalog/information_schema/table_constraints.rs
+++ b/src/frontend/src/catalog/system_catalog/information_schema/table_constraints.rs
@@ -18,7 +18,7 @@ use risingwave_frontend_macro::system_catalog;
 /// The view `table_constraints` contains all constraints belonging to tables that the current user owns or has some privilege other than SELECT on.
 /// Ref: [`https://www.postgresql.org/docs/current/infoschema-table-constraints.html`]
 /// Limitation:
-/// This view assume the constraint schema is the same as the table schema, since `pg_clatalog`.`pg_constraint` only support primrary key.
+/// This view assume the constraint schema is the same as the table schema, since `pg_catalog`.`pg_constraint` only support primary key.
 #[system_catalog(
     view,
     "information_schema.table_constraints",
diff --git a/src/frontend/src/catalog/system_catalog/pg_catalog/mod.rs b/src/frontend/src/catalog/system_catalog/pg_catalog/mod.rs
index 25fb0527cddf4..a38c8b2db441a 100644
--- a/src/frontend/src/catalog/system_catalog/pg_catalog/mod.rs
+++ b/src/frontend/src/catalog/system_catalog/pg_catalog/mod.rs
@@ -46,6 +46,7 @@ mod pg_settings;
 mod pg_shadow;
 mod pg_shdescription;
 mod pg_stat_activity;
+mod pg_stat_user_tables;
 mod pg_tables;
 mod pg_tablespace;
 mod pg_trigger;
diff --git a/src/frontend/src/catalog/system_catalog/pg_catalog/pg_stat_user_tables.rs b/src/frontend/src/catalog/system_catalog/pg_catalog/pg_stat_user_tables.rs
new file mode 100644
index 0000000000000..3aafd29c18312
--- /dev/null
+++ b/src/frontend/src/catalog/system_catalog/pg_catalog/pg_stat_user_tables.rs
@@ -0,0 +1,87 @@
+// Copyright 2024 RisingWave Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use risingwave_common::types::{Fields, Timestamptz};
+use risingwave_frontend_macro::system_catalog;
+
+/// The `pg_stat_user_tables` view will contain one row for each user table in the current database,
+/// showing statistics about accesses to that specific table.
+/// Ref: [`https://www.postgresql.org/docs/current/monitoring-stats.html#MONITORING-PG-STAT-ALL-TABLES-VIEW`]
+#[system_catalog(
+    view,
+    "pg_catalog.pg_stat_user_tables",
+    "SELECT
+        rr.id as relid,
+        rs.name as schemaname,
+        rr.name as relname,
+        NULL::bigint as seq_scan,
+        NULL::timestamptz as last_seq_scan,
+        NULL::bigint as seq_tup_read,
+        NULL::bigint as idx_scan,
+        NULL::timestamptz as last_idx_scan,
+        NULL::bigint as idx_tup_fetch,
+        NULL::bigint as n_tup_ins,
+        NULL::bigint as n_tup_del,
+        NULL::bigint as n_tup_hot_upd,
+        NULL::bigint as n_tup_newpage_upd,
+        rts.total_key_count as n_live_tup,
+        NULL::bigint as n_dead_tup,
+        NULL::bigint as n_mod_since_analyze,
+        NULL::bigint as n_ins_since_vacuum,
+        NULL::timestamptz as last_vacuum,
+        NULL::timestamptz as last_autovacuum,
+        NULL::timestamptz as last_analyze,
+        NULL::timestamptz as last_autoanalyze,
+        NULL::bigint as vacuum_count,
+        NULL::bigint as autovacuum_count,
+        NULL::bigint as analyze_count,
+        NULL::bigint as autoanalyze_count
+    FROM
+        rw_relations rr
+        left join rw_table_stats rts on rr.id = rts.id
+        join rw_schemas rs on schema_id = rs.id
+    WHERE
+        rs.name != 'rw_catalog'
+        AND rs.name != 'pg_catalog'
+        AND rs.name != 'information_schema'
+"
+)]
+#[derive(Fields)]
+struct PgStatUserTables {
+    relid: i32,
+    schemaname: String,
+    relname: String,
+    seq_scan: i64,
+    last_seq_scan: Timestamptz,
+    seq_tup_read: i64,
+    idx_scan: i64,
+    last_idx_scan: Timestamptz,
+    idx_tup_fetch: i64,
+    n_tup_ins: i64,
+    n_tup_del: i64,
+    n_tup_hot_upd: i64,
+    n_tup_newpage_upd: i64,
+    n_live_tup: i64,
+    n_dead_tup: i64,
+    n_mod_since_analyze: i64,
+    n_ins_since_vacuum: i64,
+    last_vacuum: Timestamptz,
+    last_autovacuum: Timestamptz,
+    last_analyze: Timestamptz,
+    last_autoanalyze: Timestamptz,
+    vacuum_count: i64,
+    autovacuum_count: i64,
+    analyze_count: i64,
+    autoanalyze_count: i64,
+}
diff --git a/src/frontend/src/expr/user_defined_function.rs b/src/frontend/src/expr/user_defined_function.rs
index 084fe7387d766..38d702e94b5b6 100644
--- a/src/frontend/src/expr/user_defined_function.rs
+++ b/src/frontend/src/expr/user_defined_function.rs
@@ -81,7 +81,7 @@ impl Expr for UserDefinedFunction {
         ExprNode {
             function_type: Type::Unspecified.into(),
             return_type: Some(self.return_type().to_protobuf()),
-            rex_node: Some(RexNode::Udf(UserDefinedFunction {
+            rex_node: Some(RexNode::Udf(Box::new(UserDefinedFunction {
                 children: self.args.iter().map(Expr::to_expr_proto).collect(),
                 name: self.catalog.name.clone(),
                 arg_names: self.catalog.arg_names.clone(),
@@ -98,7 +98,7 @@ impl Expr for UserDefinedFunction {
                 body: self.catalog.body.clone(),
                 compressed_binary: self.catalog.compressed_binary.clone(),
                 always_retry_on_network_error: self.catalog.always_retry_on_network_error,
-            })),
+            }))),
         }
     }
 }
diff --git a/src/frontend/src/handler/alter_table_column.rs b/src/frontend/src/handler/alter_table_column.rs
index 0342004f6b1b5..04cc9c8f8defa 100644
--- a/src/frontend/src/handler/alter_table_column.rs
+++ b/src/frontend/src/handler/alter_table_column.rs
@@ -296,10 +296,10 @@ pub(crate) fn hijack_merger_for_target_table(
         }
     }
 
-    let pb_project = PbNodeBody::Project(ProjectNode {
+    let pb_project = PbNodeBody::Project(Box::new(ProjectNode {
         select_list: exprs.iter().map(|expr| expr.to_expr_proto()).collect(),
         ..Default::default()
-    });
+    }));
 
     for fragment in graph.fragments.values_mut() {
         if let Some(node) = &mut fragment.node {
diff --git a/src/frontend/src/handler/create_sink.rs b/src/frontend/src/handler/create_sink.rs
index c89050b331d3f..3502455021ca7 100644
--- a/src/frontend/src/handler/create_sink.rs
+++ b/src/frontend/src/handler/create_sink.rs
@@ -649,9 +649,9 @@ pub(crate) fn insert_merger_to_union_with_project(
         // TODO: MergeNode is used as a placeholder, see issue #17658
         node.input.push(StreamNode {
             input: vec![StreamNode {
-                node_body: Some(NodeBody::Merge(MergeNode {
+                node_body: Some(NodeBody::Merge(Box::new(MergeNode {
                     ..Default::default()
-                })),
+                }))),
                 ..Default::default()
             }],
             identity: uniq_identity
diff --git a/src/frontend/src/handler/create_table.rs b/src/frontend/src/handler/create_table.rs
index 29af1e8c62749..1ea32fa941e6d 100644
--- a/src/frontend/src/handler/create_table.rs
+++ b/src/frontend/src/handler/create_table.rs
@@ -502,13 +502,10 @@ pub(crate) async fn gen_create_table_plan_with_source(
     format_encode: FormatEncodeOptions,
     source_watermarks: Vec<SourceWatermark>,
     mut col_id_gen: ColumnIdGenerator,
-    append_only: bool,
-    on_conflict: Option<OnConflict>,
-    with_version_column: Option<String>,
     include_column_options: IncludeOption,
-    engine: Engine,
+    props: CreateTableProps,
 ) -> Result<(PlanRef, Option<PbSource>, PbTable)> {
-    if append_only
+    if props.append_only
         && format_encode.format != Format::Plain
         && format_encode.format != Format::Native
     {
@@ -522,6 +519,9 @@ pub(crate) async fn gen_create_table_plan_with_source(
     let session = &handler_args.session;
     let with_properties = bind_connector_props(&handler_args, &format_encode, false)?;
 
+    let db_name: &str = session.database();
+    let (schema_name, _) = Binder::resolve_schema_qualified_name(db_name, table_name.clone())?;
+
     let (columns_from_resolve_source, source_info) = bind_columns_from_source(
         session,
         &format_encode,
@@ -556,14 +556,10 @@ pub(crate) async fn gen_create_table_plan_with_source(
 
     let (plan, table) = gen_table_plan_with_source(
         context.into(),
+        schema_name,
         source_catalog,
-        append_only,
-        on_conflict,
-        with_version_column,
-        Some(col_id_gen.into_version()),
-        database_id,
-        schema_id,
-        engine,
+        col_id_gen.into_version(),
+        props,
     )?;
 
     Ok((plan, Some(pb_source), table))
@@ -579,13 +575,8 @@ pub(crate) fn gen_create_table_plan(
     constraints: Vec<TableConstraint>,
     mut col_id_gen: ColumnIdGenerator,
     source_watermarks: Vec<SourceWatermark>,
-    append_only: bool,
-    on_conflict: Option<OnConflict>,
-    with_version_column: Option<String>,
-    webhook_info: Option<PbWebhookSourceInfo>,
-    engine: Engine,
+    props: CreateTableProps,
 ) -> Result<(PlanRef, PbTable)> {
-    let definition = context.normalized_sql().to_owned();
     let mut columns = bind_sql_columns(&column_defs)?;
     for c in &mut columns {
         c.column_desc.column_id = col_id_gen.generate(&*c)?;
@@ -602,14 +593,9 @@ pub(crate) fn gen_create_table_plan(
         columns,
         column_defs,
         constraints,
-        definition,
         source_watermarks,
-        append_only,
-        on_conflict,
-        with_version_column,
-        Some(col_id_gen.into_version()),
-        webhook_info,
-        engine,
+        col_id_gen.into_version(),
+        props,
     )
 }
 
@@ -620,15 +606,11 @@ pub(crate) fn gen_create_table_plan_without_source(
     columns: Vec<ColumnCatalog>,
     column_defs: Vec<ColumnDef>,
     constraints: Vec<TableConstraint>,
-    definition: String,
     source_watermarks: Vec<SourceWatermark>,
-    append_only: bool,
-    on_conflict: Option<OnConflict>,
-    with_version_column: Option<String>,
-    version: Option<TableVersion>,
-    webhook_info: Option<PbWebhookSourceInfo>,
-    engine: Engine,
+    version: TableVersion,
+    props: CreateTableProps,
 ) -> Result<(PlanRef, PbTable)> {
+    // XXX: Why not bind outside?
     let pk_names = bind_sql_pk_names(&column_defs, bind_table_constraints(&constraints)?)?;
     let (mut columns, pk_column_ids, row_id_index) =
         bind_pk_and_row_id_on_relation(columns, pk_names, true)?;
@@ -650,89 +632,97 @@ pub(crate) fn gen_create_table_plan_without_source(
     let session = context.session_ctx().clone();
 
     let db_name = session.database();
-    let (schema_name, name) = Binder::resolve_schema_qualified_name(db_name, table_name)?;
-    let (database_id, schema_id) =
-        session.get_database_and_schema_id_for_create(schema_name.clone())?;
+    let (schema_name, table_name) = Binder::resolve_schema_qualified_name(db_name, table_name)?;
 
-    gen_table_plan_inner(
-        context.into(),
-        name,
+    let info = CreateTableInfo {
         columns,
         pk_column_ids,
         row_id_index,
-        definition,
         watermark_descs,
-        append_only,
-        on_conflict,
-        with_version_column,
+        source_catalog: None,
         version,
-        None,
-        database_id,
-        schema_id,
-        webhook_info,
-        engine,
-    )
+    };
+
+    gen_table_plan_inner(context.into(), schema_name, table_name, info, props)
 }
 
 fn gen_table_plan_with_source(
     context: OptimizerContextRef,
+    schema_name: Option<String>,
     source_catalog: SourceCatalog,
-    append_only: bool,
-    on_conflict: Option<OnConflict>,
-    with_version_column: Option<String>,
-    version: Option<TableVersion>, /* TODO: this should always be `Some` if we support `ALTER
-                                    * TABLE` for `CREATE TABLE AS`. */
-    database_id: DatabaseId,
-    schema_id: SchemaId,
-    engine: Engine,
+    version: TableVersion,
+    props: CreateTableProps,
 ) -> Result<(PlanRef, PbTable)> {
-    let cloned_source_catalog = source_catalog.clone();
-    gen_table_plan_inner(
-        context,
-        source_catalog.name,
-        source_catalog.columns,
-        source_catalog.pk_col_ids,
-        source_catalog.row_id_index,
-        source_catalog.definition,
-        source_catalog.watermark_descs,
-        append_only,
-        on_conflict,
-        with_version_column,
+    let table_name = source_catalog.name.clone();
+
+    let info = CreateTableInfo {
+        columns: source_catalog.columns.clone(),
+        pk_column_ids: source_catalog.pk_col_ids.clone(),
+        row_id_index: source_catalog.row_id_index,
+        watermark_descs: source_catalog.watermark_descs.clone(),
+        source_catalog: Some(source_catalog),
         version,
-        Some(cloned_source_catalog),
-        database_id,
-        schema_id,
-        None,
-        engine,
-    )
+    };
+
+    gen_table_plan_inner(context, schema_name, table_name, info, props)
+}
+
+/// Arguments of the functions that generate a table plan, part 1.
+///
+/// Compared to [`CreateTableProps`], this struct contains fields that need some work of binding
+/// or resolving based on the user input.
+pub struct CreateTableInfo {
+    pub columns: Vec<ColumnCatalog>,
+    pub pk_column_ids: Vec<ColumnId>,
+    pub row_id_index: Option<usize>,
+    pub watermark_descs: Vec<WatermarkDesc>,
+    pub source_catalog: Option<SourceCatalog>,
+    pub version: TableVersion,
+}
+
+/// Arguments of the functions that generate a table plan, part 2.
+///
+/// Compared to [`CreateTableInfo`], this struct contains fields that can be (relatively) simply
+/// obtained from the input or the context.
+pub struct CreateTableProps {
+    pub definition: String,
+    pub append_only: bool,
+    pub on_conflict: Option<OnConflict>,
+    pub with_version_column: Option<String>,
+    pub webhook_info: Option<PbWebhookSourceInfo>,
+    pub engine: Engine,
 }
 
 #[allow(clippy::too_many_arguments)]
 fn gen_table_plan_inner(
     context: OptimizerContextRef,
+    schema_name: Option<String>,
     table_name: String,
-    columns: Vec<ColumnCatalog>,
-    pk_column_ids: Vec<ColumnId>,
-    row_id_index: Option<usize>,
-    definition: String,
-    watermark_descs: Vec<WatermarkDesc>,
-    append_only: bool,
-    on_conflict: Option<OnConflict>,
-    with_version_column: Option<String>,
-    version: Option<TableVersion>, /* TODO: this should always be `Some` if we support `ALTER
-                                    * TABLE` for `CREATE TABLE AS`. */
-    source_catalog: Option<SourceCatalog>,
-    database_id: DatabaseId,
-    schema_id: SchemaId,
-    webhook_info: Option<PbWebhookSourceInfo>,
-    engine: Engine,
+    info: CreateTableInfo,
+    props: CreateTableProps,
 ) -> Result<(PlanRef, PbTable)> {
+    let CreateTableInfo {
+        ref columns,
+        row_id_index,
+        ref watermark_descs,
+        ref source_catalog,
+        ..
+    } = info;
+    let CreateTableProps {
+        append_only,
+        on_conflict,
+        ..
+    } = props;
+
+    let (database_id, schema_id) = context
+        .session_ctx()
+        .get_database_and_schema_id_for_create(schema_name)?;
+
     let session = context.session_ctx().clone();
     let retention_seconds = context.with_options().retention_seconds();
-    let is_external_source = source_catalog.is_some();
 
     let source_node: PlanRef = LogicalSource::new(
-        source_catalog.map(|source| Rc::new(source.clone())),
+        source_catalog.clone().map(Rc::new),
         columns.clone(),
         row_id_index,
         SourceNodeKind::CreateTable,
@@ -784,20 +774,11 @@ fn gen_table_plan_inner(
     let materialize = plan_root.gen_table_plan(
         context,
         table_name,
-        columns,
-        definition,
-        pk_column_ids,
-        row_id_index,
-        append_only,
-        on_conflict,
-        with_version_column,
-        watermark_descs,
-        version,
-        is_external_source,
-        retention_seconds,
-        None,
-        webhook_info,
-        engine,
+        info,
+        CreateTableProps {
+            on_conflict,
+            ..props
+        },
     )?;
 
     let mut table = materialize.table().to_prost(schema_id, database_id);
@@ -919,24 +900,27 @@ pub(crate) fn gen_create_table_plan_for_cdc_table(
     let materialize = plan_root.gen_table_plan(
         context,
         resolved_table_name,
-        columns,
-        definition,
-        pk_column_ids,
-        None,
-        false,
-        on_conflict,
-        with_version_column,
-        vec![],
-        Some(col_id_gen.into_version()),
-        true,
-        None,
-        Some(cdc_table_id),
-        None,
-        engine,
+        CreateTableInfo {
+            columns,
+            pk_column_ids,
+            row_id_index: None,
+            watermark_descs: vec![],
+            source_catalog: Some((*source).clone()),
+            version: col_id_gen.into_version(),
+        },
+        CreateTableProps {
+            definition,
+            append_only: false,
+            on_conflict,
+            with_version_column,
+            webhook_info: None,
+            engine,
+        },
     )?;
 
     let mut table = materialize.table().to_prost(schema_id, database_id);
     table.owner = session.user_id();
+    table.cdc_table_id = Some(cdc_table_id);
     table.dependent_relations = vec![source.id];
 
     Ok((materialize.into(), table))
@@ -1024,6 +1008,18 @@ pub(super) async fn handle_create_table_plan(
         &include_column_options,
         &cdc_table_info,
     )?;
+    let webhook_info = webhook_info
+        .map(|info| bind_webhook_info(&handler_args.session, &column_defs, info))
+        .transpose()?;
+
+    let props = CreateTableProps {
+        definition: handler_args.normalized_sql.clone(),
+        append_only,
+        on_conflict,
+        with_version_column: with_version_column.clone(),
+        webhook_info,
+        engine,
+    };
 
     let ((plan, source, table), job_type) = match (format_encode, cdc_table_info.as_ref()) {
         (Some(format_encode), None) => (
@@ -1037,20 +1033,13 @@ pub(super) async fn handle_create_table_plan(
                 format_encode,
                 source_watermarks,
                 col_id_gen,
-                append_only,
-                on_conflict,
-                with_version_column,
                 include_column_options,
-                engine,
+                props,
             )
             .await?,
             TableJobType::General,
         ),
         (None, None) => {
-            let webhook_info = webhook_info
-                .map(|info| bind_webhook_info(&handler_args.session, &column_defs, info))
-                .transpose()?;
-
             let context = OptimizerContext::new(handler_args, explain_options);
             let (plan, table) = gen_create_table_plan(
                 context,
@@ -1059,11 +1048,7 @@ pub(super) async fn handle_create_table_plan(
                 constraints,
                 col_id_gen,
                 source_watermarks,
-                append_only,
-                on_conflict,
-                with_version_column,
-                webhook_info,
-                engine,
+                props,
             )?;
 
             ((plan, None, table), TableJobType::General)
@@ -1784,6 +1769,15 @@ pub async fn generate_stream_graph_for_replace_table(
 ) -> Result<(StreamFragmentGraph, Table, Option<PbSource>, TableJobType)> {
     use risingwave_pb::catalog::table::OptionalAssociatedSourceId;
 
+    let props = CreateTableProps {
+        definition: handler_args.normalized_sql.clone(),
+        append_only,
+        on_conflict,
+        with_version_column: with_version_column.clone(),
+        webhook_info: original_catalog.webhook_info.clone(),
+        engine,
+    };
+
     let ((plan, mut source, table), job_type) = match (format_encode, cdc_table_info.as_ref()) {
         (Some(format_encode), None) => (
             gen_create_table_plan_with_source(
@@ -1796,11 +1790,8 @@ pub async fn generate_stream_graph_for_replace_table(
                 format_encode,
                 source_watermarks,
                 col_id_gen,
-                append_only,
-                on_conflict,
-                with_version_column,
                 include_column_options,
-                engine,
+                props,
             )
             .await?,
             TableJobType::General,
@@ -1814,11 +1805,7 @@ pub async fn generate_stream_graph_for_replace_table(
                 constraints,
                 col_id_gen,
                 source_watermarks,
-                append_only,
-                on_conflict,
-                with_version_column,
-                original_catalog.webhook_info.clone(),
-                engine,
+                props,
             )?;
             ((plan, None, table), TableJobType::General)
         }
diff --git a/src/frontend/src/handler/create_table_as.rs b/src/frontend/src/handler/create_table_as.rs
index d90ad6afe9a73..d9019656e4e98 100644
--- a/src/frontend/src/handler/create_table_as.rs
+++ b/src/frontend/src/handler/create_table_as.rs
@@ -21,7 +21,9 @@ use risingwave_sqlparser::ast::{ColumnDef, ObjectName, OnConflict, Query, Statem
 use super::{HandlerArgs, RwPgResponse};
 use crate::binder::BoundStatement;
 use crate::error::{ErrorCode, Result};
-use crate::handler::create_table::{gen_create_table_plan_without_source, ColumnIdGenerator};
+use crate::handler::create_table::{
+    gen_create_table_plan_without_source, ColumnIdGenerator, CreateTableProps,
+};
 use crate::handler::query::handle_query;
 use crate::{build_graph, Binder, OptimizerContext};
 pub async fn handle_create_as(
@@ -107,14 +109,16 @@ pub async fn handle_create_as(
             columns,
             vec![],
             vec![],
-            "".to_owned(), // TODO: support `SHOW CREATE TABLE` for `CREATE TABLE AS`
-            vec![],        // No watermark should be defined in for `CREATE TABLE AS`
-            append_only,
-            on_conflict,
-            with_version_column,
-            Some(col_id_gen.into_version()),
-            None,
-            engine,
+            vec![], // No watermark should be defined in for `CREATE TABLE AS`
+            col_id_gen.into_version(),
+            CreateTableProps {
+                definition: "".to_owned(), // TODO: empty definition means no schema change support
+                append_only,
+                on_conflict,
+                with_version_column,
+                webhook_info: None,
+                engine,
+            },
         )?;
         let graph = build_graph(plan)?;
 
diff --git a/src/frontend/src/monitor/stats.rs b/src/frontend/src/monitor/stats.rs
index e958a4e0588e6..c3e716a8f96f8 100644
--- a/src/frontend/src/monitor/stats.rs
+++ b/src/frontend/src/monitor/stats.rs
@@ -17,7 +17,7 @@ use std::collections::HashMap;
 use std::sync::{Arc, LazyLock};
 use std::time::Duration;
 
-use prometheus::core::{AtomicI64, AtomicU64, GenericCounter, GenericGauge};
+use prometheus::core::{AtomicU64, GenericCounter};
 use prometheus::{
     exponential_buckets, histogram_opts, register_histogram_vec_with_registry,
     register_histogram_with_registry, register_int_counter_with_registry,
@@ -96,8 +96,8 @@ pub struct CursorMetrics {
     pub subscription_cursor_query_duration: HistogramVec,
     pub subscription_cursor_declare_duration: HistogramVec,
     pub subscription_cursor_fetch_duration: HistogramVec,
-    subsription_cursor_nums: GenericGauge<AtomicI64>,
-    invalid_subsription_cursor_nums: GenericGauge<AtomicI64>,
+    subsription_cursor_nums: IntGauge,
+    invalid_subsription_cursor_nums: IntGauge,
     subscription_cursor_last_fetch_duration: HistogramVec,
     _cursor_metrics_collector: Option<Arc<CursorMetricsCollector>>,
 }
@@ -199,8 +199,8 @@ struct CursorMetricsCollector {
 impl CursorMetricsCollector {
     fn new(
         session_map: SessionMapRef,
-        subsription_cursor_nums: GenericGauge<AtomicI64>,
-        invalid_subsription_cursor_nums: GenericGauge<AtomicI64>,
+        subsription_cursor_nums: IntGauge,
+        invalid_subsription_cursor_nums: IntGauge,
         subscription_cursor_last_fetch_duration: HistogramVec,
     ) -> Self {
         const COLLECT_INTERVAL_SECONDS: u64 = 60;
diff --git a/src/frontend/src/optimizer/mod.rs b/src/frontend/src/optimizer/mod.rs
index 14e45f02517f4..230224557b63b 100644
--- a/src/frontend/src/optimizer/mod.rs
+++ b/src/frontend/src/optimizer/mod.rs
@@ -51,14 +51,11 @@ pub use optimizer_context::*;
 use plan_expr_rewriter::ConstEvalRewriter;
 use property::Order;
 use risingwave_common::bail;
-use risingwave_common::catalog::{
-    ColumnCatalog, ColumnDesc, ColumnId, ConflictBehavior, Engine, Field, Schema,
-};
+use risingwave_common::catalog::{ColumnCatalog, ColumnDesc, ConflictBehavior, Field, Schema};
 use risingwave_common::types::DataType;
 use risingwave_common::util::column_index_mapping::ColIndexMapping;
 use risingwave_common::util::iter_util::ZipEqDebug;
 use risingwave_connector::sink::catalog::SinkFormatDesc;
-use risingwave_pb::catalog::{PbWebhookSourceInfo, WatermarkDesc};
 use risingwave_pb::stream_plan::StreamScanType;
 
 use self::heuristic_optimizer::ApplyOrder;
@@ -73,9 +70,10 @@ use self::plan_visitor::InputRefValidator;
 use self::plan_visitor::{has_batch_exchange, CardinalityVisitor, StreamKeyChecker};
 use self::property::{Cardinality, RequiredDist};
 use self::rule::*;
-use crate::catalog::table_catalog::{TableType, TableVersion};
+use crate::catalog::table_catalog::TableType;
 use crate::error::{ErrorCode, Result};
 use crate::expr::TimestamptzExprFinder;
+use crate::handler::create_table::{CreateTableInfo, CreateTableProps};
 use crate::optimizer::plan_node::generic::{SourceNodeKind, Union};
 use crate::optimizer::plan_node::{
     BatchExchange, PlanNodeType, PlanTreeNode, RewriteExprsRecursive, StreamExchange, StreamUnion,
@@ -639,25 +637,26 @@ impl PlanRoot {
     }
 
     /// Optimize and generate a create table plan.
-    #[allow(clippy::too_many_arguments)]
     pub fn gen_table_plan(
         mut self,
         context: OptimizerContextRef,
         table_name: String,
-        columns: Vec<ColumnCatalog>,
-        definition: String,
-        pk_column_ids: Vec<ColumnId>,
-        row_id_index: Option<usize>,
-        append_only: bool,
-        on_conflict: Option<OnConflict>,
-        with_version_column: Option<String>,
-        watermark_descs: Vec<WatermarkDesc>,
-        version: Option<TableVersion>,
-        with_external_source: bool,
-        retention_seconds: Option<NonZeroU32>,
-        cdc_table_id: Option<String>,
-        webhook_info: Option<PbWebhookSourceInfo>,
-        engine: Engine,
+        CreateTableInfo {
+            columns,
+            pk_column_ids,
+            row_id_index,
+            watermark_descs,
+            source_catalog,
+            version,
+        }: CreateTableInfo,
+        CreateTableProps {
+            definition,
+            append_only,
+            on_conflict,
+            with_version_column,
+            webhook_info,
+            engine,
+        }: CreateTableProps,
     ) -> Result<StreamMaterialize> {
         assert_eq!(self.phase, PlanPhase::Logical);
         assert_eq!(self.plan.convention(), Convention::Logical);
@@ -751,6 +750,7 @@ impl PlanRoot {
             None
         };
 
+        let with_external_source = source_catalog.is_some();
         let union_inputs = if with_external_source {
             let mut external_source_node = stream_plan;
             external_source_node =
@@ -868,6 +868,8 @@ impl PlanRoot {
             ))?
         }
 
+        let retention_seconds = context.with_options().retention_seconds();
+
         let table_required_dist = {
             let mut bitset = FixedBitSet::with_capacity(columns.len());
             for idx in &pk_column_indices {
@@ -891,7 +893,6 @@ impl PlanRoot {
             row_id_index,
             version,
             retention_seconds,
-            cdc_table_id,
             webhook_info,
             engine,
         )
diff --git a/src/frontend/src/optimizer/plan_node/stream_asof_join.rs b/src/frontend/src/optimizer/plan_node/stream_asof_join.rs
index e4ce2c7edfff2..80d2afd017f81 100644
--- a/src/frontend/src/optimizer/plan_node/stream_asof_join.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_asof_join.rs
@@ -310,7 +310,7 @@ impl StreamNode for StreamAsOfJoin {
             _ => unreachable!(),
         };
 
-        NodeBody::AsOfJoin(AsOfJoinNode {
+        NodeBody::AsOfJoin(Box::new(AsOfJoinNode {
             join_type: asof_join_type.into(),
             left_key: left_jk_indices_prost,
             right_key: right_jk_indices_prost,
@@ -321,7 +321,7 @@ impl StreamNode for StreamAsOfJoin {
             right_deduped_input_pk_indices,
             output_indices: self.core.output_indices.iter().map(|&x| x as u32).collect(),
             asof_desc: Some(self.inequality_desc),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_cdc_table_scan.rs b/src/frontend/src/optimizer/plan_node/stream_cdc_table_scan.rs
index 7ab2912bd1118..e2d57f9f5626d 100644
--- a/src/frontend/src/optimizer/plan_node/stream_cdc_table_scan.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_cdc_table_scan.rs
@@ -190,10 +190,10 @@ impl StreamCdcTableScan {
             append_only: true,
             identity: "StreamCdcFilter".to_owned(),
             fields: cdc_source_schema.clone(),
-            node_body: Some(PbNodeBody::CdcFilter(CdcFilterNode {
+            node_body: Some(PbNodeBody::CdcFilter(Box::new(CdcFilterNode {
                 search_condition: Some(filter_expr.to_expr_proto()),
                 upstream_source_id,
-            })),
+            }))),
         };
 
         let exchange_operator_id = self.core.ctx.next_plan_node_id();
@@ -205,13 +205,13 @@ impl StreamCdcTableScan {
             append_only: true,
             identity: "Exchange".to_owned(),
             fields: cdc_source_schema.clone(),
-            node_body: Some(PbNodeBody::Exchange(ExchangeNode {
+            node_body: Some(PbNodeBody::Exchange(Box::new(ExchangeNode {
                 strategy: Some(DispatchStrategy {
                     r#type: DispatcherType::Simple as _,
                     dist_key_indices: vec![], // simple exchange doesn't need dist key
                     output_indices: (0..cdc_source_schema.len() as u32).collect(),
                 }),
-            })),
+            }))),
         };
 
         // The required columns from the external table
@@ -242,7 +242,7 @@ impl StreamCdcTableScan {
         );
 
         let options = self.core.options.to_proto();
-        let stream_scan_body = PbNodeBody::StreamCdcScan(StreamCdcScanNode {
+        let stream_scan_body = PbNodeBody::StreamCdcScan(Box::new(StreamCdcScanNode {
             table_id: upstream_source_id,
             upstream_column_ids,
             output_indices,
@@ -252,7 +252,7 @@ impl StreamCdcTableScan {
             rate_limit: self.base.ctx().overwrite_options().backfill_rate_limit,
             disable_backfill: options.disable_backfill,
             options: Some(options),
-        });
+        }));
 
         // plan: merge -> filter -> exchange(simple) -> stream_scan
         Ok(PbStreamNode {
diff --git a/src/frontend/src/optimizer/plan_node/stream_changelog.rs b/src/frontend/src/optimizer/plan_node/stream_changelog.rs
index 34bfdec281815..bf09b060c4f87 100644
--- a/src/frontend/src/optimizer/plan_node/stream_changelog.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_changelog.rs
@@ -72,9 +72,9 @@ impl_distill_by_unit!(StreamChangeLog, core, "StreamChangeLog");
 
 impl StreamNode for StreamChangeLog {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
-        PbNodeBody::Changelog(ChangeLogNode {
+        PbNodeBody::Changelog(Box::new(ChangeLogNode {
             need_op: self.core.need_op,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_dedup.rs b/src/frontend/src/optimizer/plan_node/stream_dedup.rs
index d642d0f9e7ee0..689dd0432454d 100644
--- a/src/frontend/src/optimizer/plan_node/stream_dedup.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_dedup.rs
@@ -92,7 +92,7 @@ impl StreamNode for StreamDedup {
         let table_catalog = self
             .infer_internal_table_catalog()
             .with_id(state.gen_table_id_wrapped());
-        PbNodeBody::AppendOnlyDedup(DedupNode {
+        PbNodeBody::AppendOnlyDedup(Box::new(DedupNode {
             state_table: Some(table_catalog.to_internal_table_prost()),
             dedup_column_indices: self
                 .core
@@ -100,7 +100,7 @@ impl StreamNode for StreamDedup {
                 .iter()
                 .map(|idx| *idx as _)
                 .collect_vec(),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_delta_join.rs b/src/frontend/src/optimizer/plan_node/stream_delta_join.rs
index 84592aee1829a..af9694471b3ea 100644
--- a/src/frontend/src/optimizer/plan_node/stream_delta_join.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_delta_join.rs
@@ -159,7 +159,7 @@ impl TryToStreamPb for StreamDeltaJoin {
         // TODO: add a separate delta join node in proto, or move fragmenter to frontend so that we
         // don't need an intermediate representation.
         let eq_join_predicate = &self.eq_join_predicate;
-        Ok(NodeBody::DeltaIndexJoin(DeltaIndexJoinNode {
+        Ok(NodeBody::DeltaIndexJoin(Box::new(DeltaIndexJoinNode {
             join_type: self.core.join_type as i32,
             left_key: eq_join_predicate
                 .left_eq_indexes()
@@ -210,7 +210,7 @@ impl TryToStreamPb for StreamDeltaJoin {
                     .collect(),
             }),
             output_indices: self.core.output_indices.iter().map(|&x| x as u32).collect(),
-        }))
+        })))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_dml.rs b/src/frontend/src/optimizer/plan_node/stream_dml.rs
index c69baf5ad53d9..72393159a4006 100644
--- a/src/frontend/src/optimizer/plan_node/stream_dml.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_dml.rs
@@ -88,12 +88,12 @@ impl StreamNode for StreamDml {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
         use risingwave_pb::stream_plan::*;
 
-        PbNodeBody::Dml(DmlNode {
+        PbNodeBody::Dml(Box::new(DmlNode {
             table_id: 0,                                // Meta will fill this table id.
             table_version_id: INITIAL_TABLE_VERSION_ID, // Meta will fill this version id.
             column_descs: self.column_descs.iter().map(Into::into).collect(),
             rate_limit: self.base.ctx().overwrite_options().dml_rate_limit,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_dynamic_filter.rs b/src/frontend/src/optimizer/plan_node/stream_dynamic_filter.rs
index 02fb426905c7e..1f704f5c9eb74 100644
--- a/src/frontend/src/optimizer/plan_node/stream_dynamic_filter.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_dynamic_filter.rs
@@ -171,13 +171,13 @@ impl StreamNode for StreamDynamicFilter {
         let right_table = infer_right_internal_table_catalog(right.plan_base())
             .with_id(state.gen_table_id_wrapped());
         #[allow(deprecated)]
-        NodeBody::DynamicFilter(DynamicFilterNode {
+        NodeBody::DynamicFilter(Box::new(DynamicFilterNode {
             left_key: left_index as u32,
             condition,
             left_table: Some(left_table.to_internal_table_prost()),
             right_table: Some(right_table.to_internal_table_prost()),
             condition_always_relax: false, // deprecated
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_eowc_over_window.rs b/src/frontend/src/optimizer/plan_node/stream_eowc_over_window.rs
index 4d134df37799b..d27c1d417060e 100644
--- a/src/frontend/src/optimizer/plan_node/stream_eowc_over_window.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_eowc_over_window.rs
@@ -155,12 +155,12 @@ impl StreamNode for StreamEowcOverWindow {
             .with_id(state.gen_table_id_wrapped())
             .to_internal_table_prost();
 
-        PbNodeBody::EowcOverWindow(EowcOverWindowNode {
+        PbNodeBody::EowcOverWindow(Box::new(EowcOverWindowNode {
             calls,
             partition_by,
             order_by,
             state_table: Some(state_table),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_exchange.rs b/src/frontend/src/optimizer/plan_node/stream_exchange.rs
index d42f9a9392b41..28a41935d07a4 100644
--- a/src/frontend/src/optimizer/plan_node/stream_exchange.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_exchange.rs
@@ -120,7 +120,7 @@ impl_plan_tree_node_for_unary! {StreamExchange}
 
 impl StreamNode for StreamExchange {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> NodeBody {
-        NodeBody::Exchange(ExchangeNode {
+        NodeBody::Exchange(Box::new(ExchangeNode {
             strategy: if self.no_shuffle {
                 Some(DispatchStrategy {
                     r#type: DispatcherType::NoShuffle as i32,
@@ -144,7 +144,7 @@ impl StreamNode for StreamExchange {
                     output_indices: (0..self.schema().len() as u32).collect(),
                 })
             },
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_expand.rs b/src/frontend/src/optimizer/plan_node/stream_expand.rs
index fa0268a46fcf5..a6737fb22c360 100644
--- a/src/frontend/src/optimizer/plan_node/stream_expand.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_expand.rs
@@ -79,13 +79,13 @@ impl_distill_by_unit!(StreamExpand, core, "StreamExpand");
 
 impl StreamNode for StreamExpand {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
-        PbNodeBody::Expand(ExpandNode {
+        PbNodeBody::Expand(Box::new(ExpandNode {
             column_subsets: self
                 .column_subsets()
                 .iter()
                 .map(|subset| subset_to_protobuf(subset))
                 .collect(),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_filter.rs b/src/frontend/src/optimizer/plan_node/stream_filter.rs
index 0a3126ffe7180..e3e6bb90a7f0b 100644
--- a/src/frontend/src/optimizer/plan_node/stream_filter.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_filter.rs
@@ -69,9 +69,9 @@ impl_distill_by_unit!(StreamFilter, core, "StreamFilter");
 
 impl StreamNode for StreamFilter {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
-        PbNodeBody::Filter(FilterNode {
+        PbNodeBody::Filter(Box::new(FilterNode {
             search_condition: Some(ExprImpl::from(self.predicate().clone()).to_expr_proto()),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_fs_fetch.rs b/src/frontend/src/optimizer/plan_node/stream_fs_fetch.rs
index 9d3f46ca77cc3..f3ec2b280fef9 100644
--- a/src/frontend/src/optimizer/plan_node/stream_fs_fetch.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_fs_fetch.rs
@@ -123,8 +123,8 @@ impl StreamNode for StreamFsFetch {
                 secret_refs,
             }
         });
-        NodeBody::StreamFsFetch(StreamFsFetchNode {
+        NodeBody::StreamFsFetch(Box::new(StreamFsFetchNode {
             node_inner: source_inner,
-        })
+        }))
     }
 }
diff --git a/src/frontend/src/optimizer/plan_node/stream_global_approx_percentile.rs b/src/frontend/src/optimizer/plan_node/stream_global_approx_percentile.rs
index e2c795892e5f9..2b1f78f919c66 100644
--- a/src/frontend/src/optimizer/plan_node/stream_global_approx_percentile.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_global_approx_percentile.rs
@@ -132,7 +132,7 @@ impl StreamNode for StreamGlobalApproxPercentile {
                     .to_internal_table_prost(),
             ),
         };
-        PbNodeBody::GlobalApproxPercentile(body)
+        PbNodeBody::GlobalApproxPercentile(Box::new(body))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_group_topn.rs b/src/frontend/src/optimizer/plan_node/stream_group_topn.rs
index b9230270e634e..a539c0f2a1af2 100644
--- a/src/frontend/src/optimizer/plan_node/stream_group_topn.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_group_topn.rs
@@ -129,9 +129,9 @@ impl StreamNode for StreamGroupTopN {
             order_by: self.topn_order().to_protobuf(),
         };
         if self.input().append_only() {
-            PbNodeBody::AppendOnlyGroupTopN(group_topn_node)
+            PbNodeBody::AppendOnlyGroupTopN(Box::new(group_topn_node))
         } else {
-            PbNodeBody::GroupTopN(group_topn_node)
+            PbNodeBody::GroupTopN(Box::new(group_topn_node))
         }
     }
 }
diff --git a/src/frontend/src/optimizer/plan_node/stream_hash_agg.rs b/src/frontend/src/optimizer/plan_node/stream_hash_agg.rs
index 66266a893c042..542b6784069f9 100644
--- a/src/frontend/src/optimizer/plan_node/stream_hash_agg.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_hash_agg.rs
@@ -187,7 +187,7 @@ impl StreamNode for StreamHashAgg {
             self.core
                 .infer_tables(&self.base, self.vnode_col_idx, self.window_col_idx);
 
-        PbNodeBody::HashAgg(HashAggNode {
+        PbNodeBody::HashAgg(Box::new(HashAggNode {
             group_key: self.group_key().to_vec_as_u32(),
             agg_calls: self
                 .agg_calls()
@@ -220,7 +220,7 @@ impl StreamNode for StreamHashAgg {
             row_count_index: self.row_count_idx as u32,
             emit_on_window_close: self.base.emit_on_window_close(),
             version: PbAggNodeVersion::Issue13465 as _,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_hash_join.rs b/src/frontend/src/optimizer/plan_node/stream_hash_join.rs
index 0d7863a247d9c..b330a9e24c6d8 100644
--- a/src/frontend/src/optimizer/plan_node/stream_hash_join.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_hash_join.rs
@@ -359,7 +359,7 @@ impl StreamNode for StreamHashJoin {
 
         let null_safe_prost = self.eq_join_predicate.null_safes().into_iter().collect();
 
-        NodeBody::HashJoin(HashJoinNode {
+        NodeBody::HashJoin(Box::new(HashJoinNode {
             join_type: self.core.join_type as i32,
             left_key: left_jk_indices_prost,
             right_key: right_jk_indices_prost,
@@ -403,7 +403,7 @@ impl StreamNode for StreamHashJoin {
             right_deduped_input_pk_indices,
             output_indices: self.core.output_indices.iter().map(|&x| x as u32).collect(),
             is_append_only: self.is_append_only,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_hop_window.rs b/src/frontend/src/optimizer/plan_node/stream_hop_window.rs
index 4a50387c50be0..ee86b37fe35d7 100644
--- a/src/frontend/src/optimizer/plan_node/stream_hop_window.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_hop_window.rs
@@ -104,7 +104,7 @@ impl_plan_tree_node_for_unary! {StreamHopWindow}
 
 impl StreamNode for StreamHopWindow {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
-        PbNodeBody::HopWindow(HopWindowNode {
+        PbNodeBody::HopWindow(Box::new(HopWindowNode {
             time_col: self.core.time_col.index() as _,
             window_slide: Some(self.core.window_slide.into()),
             window_size: Some(self.core.window_size.into()),
@@ -121,7 +121,7 @@ impl StreamNode for StreamHopWindow {
                 .iter()
                 .map(|x| x.to_expr_proto())
                 .collect(),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_local_approx_percentile.rs b/src/frontend/src/optimizer/plan_node/stream_local_approx_percentile.rs
index b5af9be49df05..c62c1f5d45d45 100644
--- a/src/frontend/src/optimizer/plan_node/stream_local_approx_percentile.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_local_approx_percentile.rs
@@ -120,7 +120,7 @@ impl StreamNode for StreamLocalApproxPercentile {
             base,
             percentile_index,
         };
-        PbNodeBody::LocalApproxPercentile(body)
+        PbNodeBody::LocalApproxPercentile(Box::new(body))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_materialize.rs b/src/frontend/src/optimizer/plan_node/stream_materialize.rs
index 2793c84b1d555..fa912c7046985 100644
--- a/src/frontend/src/optimizer/plan_node/stream_materialize.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_materialize.rs
@@ -97,6 +97,7 @@ impl StreamMaterialize {
         } else {
             CreateType::Foreground
         };
+
         let table = Self::derive_table_catalog(
             input.clone(),
             name,
@@ -136,15 +137,14 @@ impl StreamMaterialize {
         version_column_index: Option<usize>,
         pk_column_indices: Vec<usize>,
         row_id_index: Option<usize>,
-        version: Option<TableVersion>,
+        version: TableVersion,
         retention_seconds: Option<NonZeroU32>,
-        cdc_table_id: Option<String>,
         webhook_info: Option<PbWebhookSourceInfo>,
         engine: Engine,
     ) -> Result<Self> {
         let input = Self::rewrite_input(input, user_distributed_by, TableType::Table)?;
 
-        let mut table = Self::derive_table_catalog(
+        let table = Self::derive_table_catalog(
             input.clone(),
             name,
             user_order_by,
@@ -155,7 +155,7 @@ impl StreamMaterialize {
             Some(pk_column_indices),
             row_id_index,
             TableType::Table,
-            version,
+            Some(version),
             Cardinality::unknown(), // unknown cardinality for tables
             retention_seconds,
             CreateType::Foreground,
@@ -163,8 +163,6 @@ impl StreamMaterialize {
             engine,
         )?;
 
-        table.cdc_table_id = cdc_table_id;
-
         Ok(Self::new(input, table))
     }
 
@@ -385,7 +383,7 @@ impl StreamNode for StreamMaterialize {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
         use risingwave_pb::stream_plan::*;
 
-        PbNodeBody::Materialize(MaterializeNode {
+        PbNodeBody::Materialize(Box::new(MaterializeNode {
             // We don't need table id for materialize node in frontend. The id will be generated on
             // meta catalog service.
             table_id: 0,
@@ -396,7 +394,7 @@ impl StreamNode for StreamMaterialize {
                 .map(ColumnOrder::to_protobuf)
                 .collect(),
             table: Some(self.table().to_internal_table_prost()),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_now.rs b/src/frontend/src/optimizer/plan_node/stream_now.rs
index 9ec80d15bac30..b8db60b11507b 100644
--- a/src/frontend/src/optimizer/plan_node/stream_now.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_now.rs
@@ -81,7 +81,7 @@ impl StreamNode for StreamNow {
         let table_catalog = internal_table_catalog_builder
             .build(dist_keys, 0)
             .with_id(state.gen_table_id_wrapped());
-        NodeBody::Now(PbNowNode {
+        NodeBody::Now(Box::new(PbNowNode {
             state_table: Some(table_catalog.to_internal_table_prost()),
             mode: Some(match &self.core.mode {
                 Mode::UpdateCurrent => PbNowMode::UpdateCurrent(PbNowModeUpdateCurrent {}),
@@ -93,7 +93,7 @@ impl StreamNode for StreamNow {
                     interval: Some(Datum::Some((*interval).into()).to_protobuf()),
                 }),
             }),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_over_window.rs b/src/frontend/src/optimizer/plan_node/stream_over_window.rs
index 6b0beaa9f99cc..405f25301f7b4 100644
--- a/src/frontend/src/optimizer/plan_node/stream_over_window.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_over_window.rs
@@ -130,13 +130,13 @@ impl StreamNode for StreamOverWindow {
             .config()
             .streaming_over_window_cache_policy();
 
-        PbNodeBody::OverWindow(OverWindowNode {
+        PbNodeBody::OverWindow(Box::new(OverWindowNode {
             calls,
             partition_by,
             order_by,
             state_table: Some(state_table),
             cache_policy: cache_policy.to_protobuf() as _,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_project.rs b/src/frontend/src/optimizer/plan_node/stream_project.rs
index d6ac8af7b146c..92aefd429ad9c 100644
--- a/src/frontend/src/optimizer/plan_node/stream_project.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_project.rs
@@ -159,13 +159,13 @@ impl StreamNode for StreamProject {
             .iter()
             .map(|(i, o)| (*i as u32, *o as u32))
             .unzip();
-        PbNodeBody::Project(ProjectNode {
+        PbNodeBody::Project(Box::new(ProjectNode {
             select_list: self.core.exprs.iter().map(|x| x.to_expr_proto()).collect(),
             watermark_input_cols,
             watermark_output_cols,
             nondecreasing_exprs: self.nondecreasing_exprs.iter().map(|i| *i as _).collect(),
             noop_update_hint: self.noop_update_hint,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_project_set.rs b/src/frontend/src/optimizer/plan_node/stream_project_set.rs
index 5735c4b9d5644..5e75ba7d3cb56 100644
--- a/src/frontend/src/optimizer/plan_node/stream_project_set.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_project_set.rs
@@ -111,7 +111,7 @@ impl StreamNode for StreamProjectSet {
             .iter()
             .map(|(i, o)| (*i as u32, *o as u32))
             .unzip();
-        PbNodeBody::ProjectSet(ProjectSetNode {
+        PbNodeBody::ProjectSet(Box::new(ProjectSetNode {
             select_list: self
                 .core
                 .select_list
@@ -121,7 +121,7 @@ impl StreamNode for StreamProjectSet {
             watermark_input_cols,
             watermark_expr_indices,
             nondecreasing_exprs: self.nondecreasing_exprs.iter().map(|i| *i as _).collect(),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_row_id_gen.rs b/src/frontend/src/optimizer/plan_node/stream_row_id_gen.rs
index 36b96f4dad36e..3ac868dc224fb 100644
--- a/src/frontend/src/optimizer/plan_node/stream_row_id_gen.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_row_id_gen.rs
@@ -83,9 +83,9 @@ impl StreamNode for StreamRowIdGen {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
         use risingwave_pb::stream_plan::*;
 
-        PbNodeBody::RowIdGen(RowIdGenNode {
+        PbNodeBody::RowIdGen(Box::new(RowIdGenNode {
             row_id_index: self.row_id_index as _,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_row_merge.rs b/src/frontend/src/optimizer/plan_node/stream_row_merge.rs
index 1295702135ffc..c065ffcc5a2a8 100644
--- a/src/frontend/src/optimizer/plan_node/stream_row_merge.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_row_merge.rs
@@ -141,10 +141,10 @@ impl_plan_tree_node_for_binary! { StreamRowMerge }
 
 impl StreamNode for StreamRowMerge {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
-        PbNodeBody::RowMerge(risingwave_pb::stream_plan::RowMergeNode {
+        PbNodeBody::RowMerge(Box::new(risingwave_pb::stream_plan::RowMergeNode {
             lhs_mapping: Some(self.lhs_mapping.to_protobuf()),
             rhs_mapping: Some(self.rhs_mapping.to_protobuf()),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_simple_agg.rs b/src/frontend/src/optimizer/plan_node/stream_simple_agg.rs
index f9f125654f402..66fe3424e06cf 100644
--- a/src/frontend/src/optimizer/plan_node/stream_simple_agg.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_simple_agg.rs
@@ -113,7 +113,7 @@ impl StreamNode for StreamSimpleAgg {
         let (intermediate_state_table, agg_states, distinct_dedup_tables) =
             self.core.infer_tables(&self.base, None, None);
 
-        PbNodeBody::SimpleAgg(SimpleAggNode {
+        PbNodeBody::SimpleAgg(Box::new(SimpleAggNode {
             agg_calls: self
                 .agg_calls()
                 .iter()
@@ -151,7 +151,7 @@ impl StreamNode for StreamSimpleAgg {
             row_count_index: self.row_count_idx as u32,
             version: PbAggNodeVersion::Issue13465 as _,
             must_output_per_barrier: self.must_output_per_barrier,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_sink.rs b/src/frontend/src/optimizer/plan_node/stream_sink.rs
index d9e0d116a61e5..1451a86ffb68f 100644
--- a/src/frontend/src/optimizer/plan_node/stream_sink.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_sink.rs
@@ -598,12 +598,12 @@ impl StreamNode for StreamSink {
             .infer_kv_log_store_table_catalog()
             .with_id(state.gen_table_id_wrapped());
 
-        PbNodeBody::Sink(SinkNode {
+        PbNodeBody::Sink(Box::new(SinkNode {
             sink_desc: Some(self.sink_desc.to_proto()),
             table: Some(table.to_internal_table_prost()),
             log_store_type: self.log_store_type as i32,
             rate_limit: self.base.ctx().overwrite_options().sink_rate_limit,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_sort.rs b/src/frontend/src/optimizer/plan_node/stream_sort.rs
index c4acd275f1236..3b5cbf7e5d85c 100644
--- a/src/frontend/src/optimizer/plan_node/stream_sort.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_sort.rs
@@ -128,14 +128,14 @@ impl_plan_tree_node_for_unary! { StreamEowcSort }
 impl StreamNode for StreamEowcSort {
     fn to_stream_prost_body(&self, state: &mut BuildFragmentGraphState) -> PbNodeBody {
         use risingwave_pb::stream_plan::*;
-        PbNodeBody::Sort(SortNode {
+        PbNodeBody::Sort(Box::new(SortNode {
             state_table: Some(
                 self.infer_state_table()
                     .with_id(state.gen_table_id_wrapped())
                     .to_internal_table_prost(),
             ),
             sort_column_index: self.sort_column_index as _,
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_source.rs b/src/frontend/src/optimizer/plan_node/stream_source.rs
index 909fa1e0d3009..b8e3ea6bf1834 100644
--- a/src/frontend/src/optimizer/plan_node/stream_source.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_source.rs
@@ -96,7 +96,7 @@ impl StreamNode for StreamSource {
                 secret_refs,
             }
         });
-        PbNodeBody::Source(SourceNode { source_inner })
+        PbNodeBody::Source(Box::new(SourceNode { source_inner }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_source_scan.rs b/src/frontend/src/optimizer/plan_node/stream_source_scan.rs
index f72361efbf03d..ab002a23d262e 100644
--- a/src/frontend/src/optimizer/plan_node/stream_source_scan.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_source_scan.rs
@@ -158,7 +158,7 @@ impl StreamSourceScan {
                     ..Default::default()
                 },
             ],
-            node_body: Some(PbNodeBody::SourceBackfill(backfill)),
+            node_body: Some(PbNodeBody::SourceBackfill(Box::new(backfill))),
             stream_key,
             operator_id: self.base.id().0 as u64,
             identity: self.distill_to_string(),
diff --git a/src/frontend/src/optimizer/plan_node/stream_stateless_simple_agg.rs b/src/frontend/src/optimizer/plan_node/stream_stateless_simple_agg.rs
index edb9121baf595..43ae49d3bdaa4 100644
--- a/src/frontend/src/optimizer/plan_node/stream_stateless_simple_agg.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_stateless_simple_agg.rs
@@ -84,7 +84,7 @@ impl_plan_tree_node_for_unary! { StreamStatelessSimpleAgg }
 impl StreamNode for StreamStatelessSimpleAgg {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> PbNodeBody {
         use risingwave_pb::stream_plan::*;
-        PbNodeBody::StatelessSimpleAgg(SimpleAggNode {
+        PbNodeBody::StatelessSimpleAgg(Box::new(SimpleAggNode {
             agg_calls: self
                 .agg_calls()
                 .iter()
@@ -103,7 +103,7 @@ impl StreamNode for StreamStatelessSimpleAgg {
             distinct_dedup_tables: Default::default(),
             version: AggNodeVersion::Issue13465 as _,
             must_output_per_barrier: false, // this is not used
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_table_scan.rs b/src/frontend/src/optimizer/plan_node/stream_table_scan.rs
index b75ec5b1c1d46..426d2add79f5d 100644
--- a/src/frontend/src/optimizer/plan_node/stream_table_scan.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_table_scan.rs
@@ -310,7 +310,7 @@ impl StreamTableScan {
             None
         };
 
-        let node_body = PbNodeBody::StreamScan(StreamScanNode {
+        let node_body = PbNodeBody::StreamScan(Box::new(StreamScanNode {
             table_id: self.core.table_desc.table_id.table_id,
             stream_scan_type: self.stream_scan_type as i32,
             // The column indices need to be forwarded to the downstream
@@ -322,7 +322,7 @@ impl StreamTableScan {
             arrangement_table,
             rate_limit: self.base.ctx().overwrite_options().backfill_rate_limit,
             ..Default::default()
-        });
+        }));
 
         Ok(PbStreamNode {
             fields: self.schema().to_prost(),
@@ -338,7 +338,7 @@ impl StreamTableScan {
                 },
                 // Snapshot read
                 PbStreamNode {
-                    node_body: Some(PbNodeBody::BatchPlan(batch_plan_node)),
+                    node_body: Some(PbNodeBody::BatchPlan(Box::new(batch_plan_node))),
                     operator_id: self.batch_plan_id.0 as u64,
                     identity: "BatchPlanNode".into(),
                     fields: snapshot_schema,
diff --git a/src/frontend/src/optimizer/plan_node/stream_temporal_join.rs b/src/frontend/src/optimizer/plan_node/stream_temporal_join.rs
index 390a141dcb385..b76c55a0b810c 100644
--- a/src/frontend/src/optimizer/plan_node/stream_temporal_join.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_temporal_join.rs
@@ -246,7 +246,7 @@ impl TryToStreamPb for StreamTemporalJoin {
             .as_stream_table_scan()
             .expect("should be a stream table scan");
 
-        Ok(NodeBody::TemporalJoin(TemporalJoinNode {
+        Ok(NodeBody::TemporalJoin(Box::new(TemporalJoinNode {
             join_type: self.core.join_type as i32,
             left_key: left_jk_indices_prost,
             right_key: right_jk_indices_prost,
@@ -267,7 +267,7 @@ impl TryToStreamPb for StreamTemporalJoin {
                 Some(memo_table.to_internal_table_prost())
             },
             is_nested_loop: self.is_nested_loop,
-        }))
+        })))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_topn.rs b/src/frontend/src/optimizer/plan_node/stream_topn.rs
index 80ca9141033c5..9109a76316884 100644
--- a/src/frontend/src/optimizer/plan_node/stream_topn.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_topn.rs
@@ -112,9 +112,9 @@ impl StreamNode for StreamTopN {
             order_by: self.topn_order().to_protobuf(),
         };
         if self.input().append_only() {
-            PbNodeBody::AppendOnlyTopN(topn_node)
+            PbNodeBody::AppendOnlyTopN(Box::new(topn_node))
         } else {
-            PbNodeBody::TopN(topn_node)
+            PbNodeBody::TopN(Box::new(topn_node))
         }
     }
 }
diff --git a/src/frontend/src/optimizer/plan_node/stream_values.rs b/src/frontend/src/optimizer/plan_node/stream_values.rs
index 0a71c208c32ee..eb42ecba57852 100644
--- a/src/frontend/src/optimizer/plan_node/stream_values.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_values.rs
@@ -72,7 +72,7 @@ impl Distill for StreamValues {
 
 impl StreamNode for StreamValues {
     fn to_stream_prost_body(&self, _state: &mut BuildFragmentGraphState) -> ProstStreamNode {
-        ProstStreamNode::Values(ValuesNode {
+        ProstStreamNode::Values(Box::new(ValuesNode {
             tuples: self
                 .logical
                 .rows()
@@ -86,7 +86,7 @@ impl StreamNode for StreamValues {
                 .iter()
                 .map(|f| f.to_prost())
                 .collect(),
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/optimizer/plan_node/stream_watermark_filter.rs b/src/frontend/src/optimizer/plan_node/stream_watermark_filter.rs
index 63699998ea4b0..1d0935225abb0 100644
--- a/src/frontend/src/optimizer/plan_node/stream_watermark_filter.rs
+++ b/src/frontend/src/optimizer/plan_node/stream_watermark_filter.rs
@@ -148,12 +148,12 @@ impl StreamNode for StreamWatermarkFilter {
 
         let table = infer_internal_table_catalog(watermark_type);
 
-        PbNodeBody::WatermarkFilter(WatermarkFilterNode {
+        PbNodeBody::WatermarkFilter(Box::new(WatermarkFilterNode {
             watermark_descs: self.watermark_descs.clone(),
             tables: vec![table
                 .with_id(state.gen_table_id_wrapped())
                 .to_internal_table_prost()],
-        })
+        }))
     }
 }
 
diff --git a/src/frontend/src/scheduler/distributed/stats.rs b/src/frontend/src/scheduler/distributed/stats.rs
index 0cda386c6de88..506426f7fa144 100644
--- a/src/frontend/src/scheduler/distributed/stats.rs
+++ b/src/frontend/src/scheduler/distributed/stats.rs
@@ -14,16 +14,17 @@
 
 use std::sync::LazyLock;
 
-use prometheus::core::{AtomicI64, AtomicU64, GenericCounter, GenericGauge};
+use prometheus::core::{AtomicU64, GenericCounter};
 use prometheus::{
     exponential_buckets, histogram_opts, register_histogram_with_registry,
-    register_int_counter_with_registry, register_int_gauge_with_registry, Histogram, Registry,
+    register_int_counter_with_registry, register_int_gauge_with_registry, Histogram, IntGauge,
+    Registry,
 };
 use risingwave_common::monitor::GLOBAL_METRICS_REGISTRY;
 
 #[derive(Clone)]
 pub struct DistributedQueryMetrics {
-    pub running_query_num: GenericGauge<AtomicI64>,
+    pub running_query_num: IntGauge,
     pub rejected_query_counter: GenericCounter<AtomicU64>,
     pub completed_query_counter: GenericCounter<AtomicU64>,
     pub query_latency: Histogram,
diff --git a/src/frontend/src/stream_fragmenter/mod.rs b/src/frontend/src/stream_fragmenter/mod.rs
index f30b0abf5b4c4..b9447fd678444 100644
--- a/src/frontend/src/stream_fragmenter/mod.rs
+++ b/src/frontend/src/stream_fragmenter/mod.rs
@@ -439,9 +439,9 @@ fn build_fragment(
                                 let node = state.gen_no_op_stream_node(StreamNode {
                                     operator_id: no_shuffle_exchange_operator_id,
                                     identity: "StreamNoShuffleExchange".into(),
-                                    node_body: Some(NodeBody::Exchange(ExchangeNode {
+                                    node_body: Some(NodeBody::Exchange(Box::new(ExchangeNode {
                                         strategy: Some(no_shuffle_strategy.clone()),
-                                    })),
+                                    }))),
                                     input: vec![],
 
                                     // Take reference's properties.
diff --git a/src/frontend/src/stream_fragmenter/rewrite/delta_join.rs b/src/frontend/src/stream_fragmenter/rewrite/delta_join.rs
index 8b6a2cb06c871..5c441a3aefeab 100644
--- a/src/frontend/src/stream_fragmenter/rewrite/delta_join.rs
+++ b/src/frontend/src/stream_fragmenter/rewrite/delta_join.rs
@@ -35,11 +35,11 @@ fn build_no_shuffle_exchange_for_delta_join(
         identity: "NO SHUFFLE Exchange (Lookup and Merge)".into(),
         fields: upstream.fields.clone(),
         stream_key: upstream.stream_key.clone(),
-        node_body: Some(NodeBody::Exchange(ExchangeNode {
+        node_body: Some(NodeBody::Exchange(Box::new(ExchangeNode {
             strategy: Some(dispatch_no_shuffle(
                 (0..(upstream.fields.len() as u32)).collect(),
             )),
-        })),
+        }))),
         input: vec![],
         append_only: upstream.append_only,
     }
@@ -55,12 +55,12 @@ fn build_consistent_hash_shuffle_exchange_for_delta_join(
         identity: "HASH Exchange (Lookup and Merge)".into(),
         fields: upstream.fields.clone(),
         stream_key: upstream.stream_key.clone(),
-        node_body: Some(NodeBody::Exchange(ExchangeNode {
+        node_body: Some(NodeBody::Exchange(Box::new(ExchangeNode {
             strategy: Some(dispatch_consistent_hash_shuffle(
                 dist_key_indices,
                 (0..(upstream.fields.len() as u32)).collect(),
             )),
-        })),
+        }))),
         input: vec![],
         append_only: upstream.append_only,
     }
@@ -97,7 +97,7 @@ fn build_lookup_for_delta_join(
         identity: "Lookup".into(),
         fields: output_fields,
         stream_key: output_stream_key,
-        node_body: Some(NodeBody::Lookup(lookup_node)),
+        node_body: Some(NodeBody::Lookup(Box::new(lookup_node))),
         input: vec![
             exchange_node_arrangement.clone(),
             exchange_node_stream.clone(),
@@ -281,7 +281,9 @@ fn build_delta_join_inner(
         identity: "Union".into(),
         fields: node.fields.clone(),
         stream_key: node.stream_key.clone(),
-        node_body: Some(NodeBody::LookupUnion(LookupUnionNode { order: vec![1, 0] })),
+        node_body: Some(NodeBody::LookupUnion(Box::new(LookupUnionNode {
+            order: vec![1, 0],
+        }))),
         input: vec![exchange_l0m.clone(), exchange_l1m.clone()],
         append_only: node.append_only,
     };
diff --git a/src/frontend/src/utils/condition.rs b/src/frontend/src/utils/condition.rs
index e497ba821864a..fb5fd07c1fc07 100644
--- a/src/frontend/src/utils/condition.rs
+++ b/src/frontend/src/utils/condition.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::cmp::Ordering;
 use std::collections::{BTreeMap, HashSet};
 use std::fmt::{self, Debug};
 use std::ops::Bound;
@@ -24,6 +25,7 @@ use risingwave_common::catalog::{Schema, TableDesc};
 use risingwave_common::types::{DataType, DefaultOrd, ScalarImpl};
 use risingwave_common::util::iter_util::ZipEqFast;
 use risingwave_common::util::scan_range::{is_full_range, ScanRange};
+use risingwave_common::util::sort_util::{cmp_rows, OrderType};
 
 use crate::error::Result;
 use crate::expr::{
@@ -299,7 +301,7 @@ impl Condition {
         table_desc: Rc<TableDesc>,
         max_split_range_gap: u64,
         disjunctions: Vec<ExprImpl>,
-    ) -> Result<Option<(Vec<ScanRange>, Self)>> {
+    ) -> Result<Option<(Vec<ScanRange>, bool)>> {
         let disjunctions_result: Result<Vec<(Vec<ScanRange>, Self)>> = disjunctions
             .into_iter()
             .map(|x| {
@@ -352,9 +354,154 @@ impl Condition {
                 }
             }
 
-            Ok(Some((non_overlap_scan_ranges, Condition::true_cond())))
+            Ok(Some((non_overlap_scan_ranges, false)))
         } else {
-            Ok(None)
+            let mut scan_ranges = vec![];
+            for (scan_ranges_chunk, _) in disjunctions_result {
+                if scan_ranges_chunk.is_empty() {
+                    // full scan range
+                    return Ok(None);
+                }
+
+                scan_ranges.extend(scan_ranges_chunk);
+            }
+
+            let order_types = table_desc
+                .pk
+                .iter()
+                .cloned()
+                .map(|x| {
+                    if x.order_type.is_descending() {
+                        x.order_type.reverse()
+                    } else {
+                        x.order_type
+                    }
+                })
+                .collect_vec();
+            scan_ranges.sort_by(|left, right| {
+                let (left_start, _left_end) = &left.convert_to_range();
+                let (right_start, _right_end) = &right.convert_to_range();
+
+                let left_start_vec = match &left_start {
+                    Bound::Included(vec) | Bound::Excluded(vec) => vec,
+                    _ => &vec![],
+                };
+                let right_start_vec = match &right_start {
+                    Bound::Included(vec) | Bound::Excluded(vec) => vec,
+                    _ => &vec![],
+                };
+
+                if left_start_vec.is_empty() && right_start_vec.is_empty() {
+                    return Ordering::Less;
+                }
+
+                if left_start_vec.is_empty() {
+                    return Ordering::Less;
+                }
+
+                if right_start_vec.is_empty() {
+                    return Ordering::Greater;
+                }
+
+                let cmp_column_len = left_start_vec.len().min(right_start_vec.len());
+                cmp_rows(
+                    &left_start_vec[0..cmp_column_len],
+                    &right_start_vec[0..cmp_column_len],
+                    &order_types[0..cmp_column_len],
+                )
+            });
+
+            if scan_ranges.is_empty() {
+                return Ok(None);
+            }
+
+            if scan_ranges.len() == 1 {
+                return Ok(Some((scan_ranges, true)));
+            }
+
+            let mut output_scan_ranges: Vec<ScanRange> = vec![];
+            output_scan_ranges.push(scan_ranges[0].clone());
+            let mut idx = 1;
+            loop {
+                if idx >= scan_ranges.len() {
+                    break;
+                }
+
+                let scan_range_left = output_scan_ranges.last_mut().unwrap();
+                let scan_range_right = &scan_ranges[idx];
+
+                if scan_range_left.eq_conds == scan_range_right.eq_conds {
+                    // range merge
+
+                    if !ScanRange::is_overlap(scan_range_left, scan_range_right, &order_types) {
+                        // not merge
+                        output_scan_ranges.push(scan_range_right.clone());
+                        idx += 1;
+                        continue;
+                    }
+
+                    // merge range
+                    fn merge_bound(
+                        left_scan_range: &Bound<Vec<Option<ScalarImpl>>>,
+                        right_scan_range: &Bound<Vec<Option<ScalarImpl>>>,
+                        order_types: &[OrderType],
+                        left_bound: bool,
+                    ) -> Bound<Vec<Option<ScalarImpl>>> {
+                        let left_scan_range = match left_scan_range {
+                            Bound::Included(vec) | Bound::Excluded(vec) => vec,
+                            Bound::Unbounded => return Bound::Unbounded,
+                        };
+
+                        let right_scan_range = match right_scan_range {
+                            Bound::Included(vec) | Bound::Excluded(vec) => vec,
+                            Bound::Unbounded => return Bound::Unbounded,
+                        };
+
+                        let cmp_len = left_scan_range.len().min(right_scan_range.len());
+
+                        let cmp = cmp_rows(
+                            &left_scan_range[..cmp_len],
+                            &right_scan_range[..cmp_len],
+                            &order_types[..cmp_len],
+                        );
+
+                        let bound = {
+                            if (cmp.is_le() && left_bound) || (cmp.is_ge() && !left_bound) {
+                                left_scan_range.to_vec()
+                            } else {
+                                right_scan_range.to_vec()
+                            }
+                        };
+
+                        // Included Bound just for convenience, the correctness will be guaranteed by the upper level filter.
+                        Bound::Included(bound)
+                    }
+
+                    scan_range_left.range.0 = merge_bound(
+                        &scan_range_left.range.0,
+                        &scan_range_right.range.0,
+                        &order_types,
+                        true,
+                    );
+
+                    scan_range_left.range.1 = merge_bound(
+                        &scan_range_left.range.1,
+                        &scan_range_right.range.1,
+                        &order_types,
+                        false,
+                    );
+
+                    if scan_range_left.is_full_table_scan() {
+                        return Ok(None);
+                    }
+                } else {
+                    output_scan_ranges.push(scan_range_right.clone());
+                }
+
+                idx += 1;
+            }
+
+            Ok(Some((output_scan_ranges, true)))
         }
     }
 
@@ -415,28 +562,43 @@ impl Condition {
                     || matches!(func_type, ExprType::GreaterThan))
             {
                 let mut pk_struct = vec![];
+                let mut order_type = None;
                 let mut all_added = true;
                 let mut iter = row_left_inputs.iter().zip_eq_fast(right_iter);
-                for i in 0..table_desc.order_column_indices().len() {
+                for column_order in &table_desc.pk {
                     if let Some((left_expr, right_expr)) = iter.next() {
-                        if left_expr.as_input_ref().unwrap().index != i {
+                        if left_expr.as_input_ref().unwrap().index != column_order.column_index {
                             all_added = false;
                             break;
                         }
+                        match order_type {
+                            Some(o) => {
+                                if o != column_order.order_type {
+                                    all_added = false;
+                                    break;
+                                }
+                            }
+                            None => order_type = Some(column_order.order_type),
+                        }
                         pk_struct.push(right_expr.clone());
                     }
                 }
 
+                // Here it is necessary to determine whether all of row is included in the `ScanRanges`, if so, the data for eq is not needed
                 if !pk_struct.is_empty() {
-                    let scan_range = ScanRange {
-                        eq_conds: vec![],
-                        range: match func_type {
-                            ExprType::GreaterThan => (Bound::Excluded(pk_struct), Bound::Unbounded),
-                            ExprType::LessThan => (Bound::Unbounded, Bound::Excluded(pk_struct)),
-                            _ => unreachable!(),
-                        },
-                    };
                     if !all_added {
+                        let scan_range = ScanRange {
+                            eq_conds: vec![],
+                            range: match func_type {
+                                ExprType::GreaterThan => {
+                                    (Bound::Included(pk_struct), Bound::Unbounded)
+                                }
+                                ExprType::LessThan => {
+                                    (Bound::Unbounded, Bound::Included(pk_struct))
+                                }
+                                _ => unreachable!(),
+                            },
+                        };
                         return Ok(Some((
                             vec![scan_range],
                             Condition {
@@ -444,6 +606,18 @@ impl Condition {
                             },
                         )));
                     } else {
+                        let scan_range = ScanRange {
+                            eq_conds: vec![],
+                            range: match func_type {
+                                ExprType::GreaterThan => {
+                                    (Bound::Excluded(pk_struct), Bound::Unbounded)
+                                }
+                                ExprType::LessThan => {
+                                    (Bound::Unbounded, Bound::Excluded(pk_struct))
+                                }
+                                _ => unreachable!(),
+                            },
+                        };
                         return Ok(Some((
                             vec![scan_range],
                             Condition {
@@ -470,12 +644,18 @@ impl Condition {
         // It's an OR.
         if self.conjunctions.len() == 1 {
             if let Some(disjunctions) = self.conjunctions[0].as_or_disjunctions() {
-                if let Some((scan_ranges, other_condition)) = Self::disjunctions_to_scan_ranges(
-                    table_desc,
-                    max_split_range_gap,
-                    disjunctions,
-                )? {
-                    return Ok((scan_ranges, other_condition));
+                if let Some((scan_ranges, maintaining_condition)) =
+                    Self::disjunctions_to_scan_ranges(
+                        table_desc,
+                        max_split_range_gap,
+                        disjunctions,
+                    )?
+                {
+                    if maintaining_condition {
+                        return Ok((scan_ranges, self));
+                    } else {
+                        return Ok((scan_ranges, Condition::true_cond()));
+                    }
                 } else {
                     return Ok((vec![], self));
                 }
diff --git a/src/java_binding/Cargo.toml b/src/java_binding/Cargo.toml
index 0966b700a713f..36facf1790abd 100644
--- a/src/java_binding/Cargo.toml
+++ b/src/java_binding/Cargo.toml
@@ -3,12 +3,6 @@ name = "risingwave_java_binding"
 version = "0.1.0"
 edition = "2021"
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 bytes = "1"
diff --git a/src/jni_core/Cargo.toml b/src/jni_core/Cargo.toml
index 868c79e7427ec..d182e19045997 100644
--- a/src/jni_core/Cargo.toml
+++ b/src/jni_core/Cargo.toml
@@ -3,29 +3,18 @@ name = "risingwave_jni_core"
 version = "0.1.0"
 edition = "2021"
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 bytes = "1"
 cfg-or-panic = "0.2"
 chrono = { version = "0.4", default-features = false }
-foyer ={ workspace = true }
 fs-err = "3"
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
-itertools = { workspace = true }
 jni = { version = "0.21.1", features = ["invocation"] }
 paste = "1"
 prost = { workspace = true }
 risingwave_common = { workspace = true }
 risingwave_pb = { workspace = true }
-rw_futures_util = { workspace = true }
-serde = { version = "1.0", features = ["derive"] }
-serde_json = "1.0"
 thiserror = "1"
 thiserror-ext = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio", features = [
@@ -41,8 +30,6 @@ tracing = "0.1"
 
 [dev-dependencies]
 expect-test = "1"
-risingwave_expr = { workspace = true }
-risingwave_hummock_sdk = { workspace = true }
 
 [lints]
 workspace = true
diff --git a/src/license/Cargo.toml b/src/license/Cargo.toml
index b435747467e21..3c0c2e8f612ff 100644
--- a/src/license/Cargo.toml
+++ b/src/license/Cargo.toml
@@ -8,12 +8,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 jsonbb = { workspace = true }
 jsonwebtoken = "9"
diff --git a/src/meta/Cargo.toml b/src/meta/Cargo.toml
index 16a303609699b..582903c99a00a 100644
--- a/src/meta/Cargo.toml
+++ b/src/meta/Cargo.toml
@@ -7,21 +7,12 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 arc-swap = "1"
 assert_matches = "1"
 async-trait = "0.1"
-aws-config = { workspace = true }
-aws-sdk-ec2 = { workspace = true }
 base64-url = { version = "3.0.0" }
-bincode = "1.3"
 bytes = { version = "1", features = ["serde"] }
 chrono = "0.4"
 clap = { workspace = true }
@@ -32,17 +23,12 @@ educe = "0.6"
 either = "1"
 enum-as-inner = "0.6"
 fail = "0.5"
-flate2 = "1"
-function_name = "0.3.0"
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
 hex = "0.4"
 http = "1"
 itertools = { workspace = true }
 jsonbb = { workspace = true }
 maplit = "1.0.2"
-memcomparable = { version = "0.2" }
-mime_guess = "2"
-moka = { version = "0.12.3", features = ["future"] }
 notify = { version = "7", default-features = false, features = [
     "macos_fsevent",
 ] }
@@ -89,7 +75,6 @@ tokio-stream = { workspace = true }
 tonic = { workspace = true }
 tower = { version = "0.5", features = ["util", "load-shed"] }
 tracing = "0.1"
-url = "2"
 uuid = { version = "1", features = ["v4"] }
 
 [target.'cfg(not(madsim))'.dependencies]
diff --git a/src/meta/model/Cargo.toml b/src/meta/model/Cargo.toml
index 991becc820642..530a67da3b9f5 100644
--- a/src/meta/model/Cargo.toml
+++ b/src/meta/model/Cargo.toml
@@ -7,17 +7,10 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 prost = { workspace = true }
 risingwave_common = { workspace = true }
-risingwave_hummock_sdk = { workspace = true }
 risingwave_pb = { workspace = true }
 sea-orm = { workspace = true }
 serde = { version = "1.0.196", features = ["derive"] }
-serde_json = "1.0.113"
+serde_json = "1"
diff --git a/src/meta/model/migration/Cargo.toml b/src/meta/model/migration/Cargo.toml
index 9d9fbd5292d25..725e028cabdc7 100644
--- a/src/meta/model/migration/Cargo.toml
+++ b/src/meta/model/migration/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 async-std = { version = "1", features = ["attributes", "tokio1"] }
 easy-ext = "1"
diff --git a/src/meta/node/Cargo.toml b/src/meta/node/Cargo.toml
index dc011842c3cd3..f3f13d546ec06 100644
--- a/src/meta/node/Cargo.toml
+++ b/src/meta/node/Cargo.toml
@@ -7,20 +7,10 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
-anyhow = "1"
 clap = { workspace = true }
 educe = "0.6"
-either = "1"
-futures = { version = "0.3", default-features = false, features = ["alloc"] }
 hex = "0.4"
-itertools = { workspace = true }
 otlp-embedded = { workspace = true }
 prometheus-http-query = "0.8"
 redact = "0.1.5"
@@ -29,12 +19,10 @@ risingwave_common = { workspace = true }
 risingwave_common_heap_profiling = { workspace = true }
 risingwave_common_service = { workspace = true }
 risingwave_meta = { workspace = true }
-risingwave_meta_model_migration = { workspace = true }
 risingwave_meta_service = { workspace = true }
 risingwave_pb = { workspace = true }
 risingwave_rpc_client = { workspace = true }
 sea-orm = { workspace = true }
-serde = { version = "1", features = ["derive"] }
 serde_json = "1"
 thiserror-ext = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio", features = [
diff --git a/src/meta/service/Cargo.toml b/src/meta/service/Cargo.toml
index 53c3708da0e12..b3fbe5e8bcb78 100644
--- a/src/meta/service/Cargo.toml
+++ b/src/meta/service/Cargo.toml
@@ -7,19 +7,11 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 async-trait = "0.1"
-either = "1"
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
 itertools = { workspace = true }
-prost = { workspace = true }
 rand = { workspace = true }
 regex = "1"
 risingwave_common = { workspace = true }
@@ -30,7 +22,6 @@ risingwave_meta_model = { workspace = true }
 risingwave_pb = { workspace = true }
 sea-orm = { workspace = true }
 serde_json = "1"
-sync-point = { path = "../../utils/sync-point" }
 thiserror-ext = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio", features = [
     "rt",
diff --git a/src/meta/src/barrier/checkpoint/control.rs b/src/meta/src/barrier/checkpoint/control.rs
index 28ee5aa0bfe32..4a22ff3a0ac2f 100644
--- a/src/meta/src/barrier/checkpoint/control.rs
+++ b/src/meta/src/barrier/checkpoint/control.rs
@@ -50,6 +50,7 @@ use crate::barrier::{
 };
 use crate::manager::ActiveStreamingWorkerNodes;
 use crate::rpc::metrics::GLOBAL_META_METRICS;
+use crate::stream::fill_snapshot_backfill_epoch;
 use crate::{MetaError, MetaResult};
 
 #[derive(Default)]
@@ -695,8 +696,10 @@ impl DatabaseCheckpointControl {
                                 progress_epoch,
                                 creating_job
                                     .snapshot_backfill_info
-                                    .upstream_mv_table_ids
-                                    .clone(),
+                                    .upstream_mv_table_id_to_backfill_epoch
+                                    .keys()
+                                    .cloned()
+                                    .collect(),
                             ),
                         )
                     })
@@ -977,7 +980,7 @@ impl DatabaseCheckpointControl {
         if let Some(Command::CreateStreamingJob {
             job_type: CreateStreamingJobType::SnapshotBackfill(snapshot_backfill_info),
             info,
-        }) = &command
+        }) = &mut command
         {
             if self.state.paused_reason().is_some() {
                 warn!("cannot create streaming job with snapshot backfill when paused");
@@ -989,18 +992,44 @@ impl DatabaseCheckpointControl {
                 }
                 return Ok(());
             }
+            // set snapshot epoch of upstream table for snapshot backfill
+            for snapshot_backfill_epoch in snapshot_backfill_info
+                .upstream_mv_table_id_to_backfill_epoch
+                .values_mut()
+            {
+                assert!(
+                    snapshot_backfill_epoch
+                        .replace(barrier_info.prev_epoch())
+                        .is_none(),
+                    "must not set previously"
+                );
+            }
+            for stream_actor in info
+                .stream_job_fragments
+                .fragments
+                .values_mut()
+                .flat_map(|fragment| fragment.actors.iter_mut())
+            {
+                fill_snapshot_backfill_epoch(
+                    stream_actor.nodes.as_mut().expect("should exist"),
+                    &snapshot_backfill_info.upstream_mv_table_id_to_backfill_epoch,
+                )?;
+            }
+            let info = info.clone();
+            let job_id = info.stream_job_fragments.stream_job_id();
+            let snapshot_backfill_info = snapshot_backfill_info.clone();
             let mutation = command
                 .as_ref()
                 .expect("checked Some")
                 .to_mutation(None)
                 .expect("should have some mutation in `CreateStreamingJob` command");
-            let job_id = info.stream_job_fragments.stream_job_id();
+
             control_stream_manager.add_partial_graph(self.database_id, Some(job_id))?;
             self.creating_streaming_job_controls.insert(
                 job_id,
                 CreatingStreamingJobControl::new(
-                    info.clone(),
-                    snapshot_backfill_info.clone(),
+                    info,
+                    snapshot_backfill_info,
                     barrier_info.prev_epoch(),
                     hummock_version_stats,
                     mutation,
diff --git a/src/meta/src/barrier/command.rs b/src/meta/src/barrier/command.rs
index 3ded5c438a670..b42986d354a60 100644
--- a/src/meta/src/barrier/command.rs
+++ b/src/meta/src/barrier/command.rs
@@ -227,7 +227,10 @@ impl CreateStreamingJobCommandInfo {
 
 #[derive(Debug, Clone)]
 pub struct SnapshotBackfillInfo {
-    pub upstream_mv_table_ids: HashSet<TableId>,
+    /// `table_id` -> `Some(snapshot_backfill_epoch)`
+    /// The `snapshot_backfill_epoch` should be None at the beginning, and be filled
+    /// by global barrier worker when handling the command.
+    pub upstream_mv_table_id_to_backfill_epoch: HashMap<TableId, Option<u64>>,
 }
 
 #[derive(Debug, Clone)]
@@ -697,8 +700,8 @@ impl Command {
                             job_type
                         {
                             snapshot_backfill_info
-                                .upstream_mv_table_ids
-                                .iter()
+                                .upstream_mv_table_id_to_backfill_epoch
+                                .keys()
                                 .map(|table_id| SubscriptionUpstreamInfo {
                                     subscriber_id: table_fragments.stream_job_id().table_id,
                                     upstream_mv_table_id: table_id.table_id,
@@ -747,12 +750,13 @@ impl Command {
                         info: jobs_to_merge
                             .iter()
                             .flat_map(|(table_id, (backfill_info, _))| {
-                                backfill_info.upstream_mv_table_ids.iter().map(
-                                    move |upstream_table_id| SubscriptionUpstreamInfo {
+                                backfill_info
+                                    .upstream_mv_table_id_to_backfill_epoch
+                                    .keys()
+                                    .map(move |upstream_table_id| SubscriptionUpstreamInfo {
                                         subscriber_id: table_id.table_id,
                                         upstream_mv_table_id: upstream_table_id.table_id,
-                                    },
-                                )
+                                    })
                             })
                             .collect(),
                     }))
diff --git a/src/meta/src/barrier/context/context_impl.rs b/src/meta/src/barrier/context/context_impl.rs
index f1306a99d7823..a045708b72dd4 100644
--- a/src/meta/src/barrier/context/context_impl.rs
+++ b/src/meta/src/barrier/context/context_impl.rs
@@ -19,6 +19,7 @@ use risingwave_common::catalog::DatabaseId;
 use risingwave_pb::common::WorkerNode;
 use risingwave_pb::hummock::HummockVersionStats;
 use risingwave_pb::meta::PausedReason;
+use risingwave_pb::stream_plan::PbFragmentTypeFlag;
 use risingwave_pb::stream_service::streaming_control_stream_request::PbInitRequest;
 use risingwave_pb::stream_service::WaitEpochCommitRequest;
 use risingwave_rpc_client::StreamingControlHandle;
@@ -178,8 +179,57 @@ impl CommandContext {
                     .unregister_table_ids(unregistered_state_table_ids.iter().cloned())
                     .await?;
             }
-
             Command::CreateStreamingJob { info, job_type } => {
+                let mut fragment_replacements = None;
+                let mut dropped_actors = None;
+                match job_type {
+                    CreateStreamingJobType::SinkIntoTable(
+                        replace_plan @ ReplaceStreamJobPlan {
+                            new_fragments,
+                            dispatchers,
+                            init_split_assignment,
+                            ..
+                        },
+                    ) => {
+                        barrier_manager_context
+                            .metadata_manager
+                            .catalog_controller
+                            .post_collect_job_fragments(
+                                new_fragments.stream_job_id().table_id as _,
+                                new_fragments.actor_ids(),
+                                dispatchers.clone(),
+                                init_split_assignment,
+                            )
+                            .await?;
+                        fragment_replacements = Some(replace_plan.fragment_replacements());
+                        dropped_actors = Some(replace_plan.dropped_actors());
+                    }
+                    CreateStreamingJobType::Normal => {}
+                    CreateStreamingJobType::SnapshotBackfill(snapshot_backfill_info) => {
+                        barrier_manager_context
+                            .metadata_manager
+                            .catalog_controller
+                            .fill_snapshot_backfill_epoch(
+                                info.stream_job_fragments.fragments.iter().filter_map(
+                                    |(fragment_id, fragment)| {
+                                        if (fragment.fragment_type_mask
+                                            & PbFragmentTypeFlag::SnapshotBackfillStreamScan as u32)
+                                            != 0
+                                        {
+                                            Some(*fragment_id as _)
+                                        } else {
+                                            None
+                                        }
+                                    },
+                                ),
+                                &snapshot_backfill_info.upstream_mv_table_id_to_backfill_epoch,
+                            )
+                            .await?
+                    }
+                }
+
+                // Do `post_collect_job_fragments` of the original streaming job in the end, so that in any previous failure,
+                // we won't mark the job as `Creating`, and then the job will be later clean by the recovery triggered by the returned error.
                 let CreateStreamingJobCommandInfo {
                     stream_job_fragments,
                     dispatchers,
@@ -197,31 +247,6 @@ impl CommandContext {
                     )
                     .await?;
 
-                let mut fragment_replacements = None;
-                let mut dropped_actors = None;
-                if let CreateStreamingJobType::SinkIntoTable(
-                    replace_plan @ ReplaceStreamJobPlan {
-                        new_fragments,
-                        dispatchers,
-                        init_split_assignment,
-                        ..
-                    },
-                ) = job_type
-                {
-                    barrier_manager_context
-                        .metadata_manager
-                        .catalog_controller
-                        .post_collect_job_fragments(
-                            new_fragments.stream_job_id().table_id as _,
-                            new_fragments.actor_ids(),
-                            dispatchers.clone(),
-                            init_split_assignment,
-                        )
-                        .await?;
-                    fragment_replacements = Some(replace_plan.fragment_replacements());
-                    dropped_actors = Some(replace_plan.dropped_actors());
-                }
-
                 // Extract the fragments that include source operators.
                 let source_fragments = stream_job_fragments.stream_source_fragments();
                 let backfill_fragments = stream_job_fragments.source_backfill_fragments()?;
diff --git a/src/meta/src/controller/fragment.rs b/src/meta/src/controller/fragment.rs
index a9a536f0b394d..6e71835618dfb 100644
--- a/src/meta/src/controller/fragment.rs
+++ b/src/meta/src/controller/fragment.rs
@@ -1293,6 +1293,34 @@ impl CatalogController {
         Ok(())
     }
 
+    pub async fn fill_snapshot_backfill_epoch(
+        &self,
+        fragment_ids: impl Iterator<Item = FragmentId>,
+        upstream_mv_snapshot_epoch: &HashMap<risingwave_common::catalog::TableId, Option<u64>>,
+    ) -> MetaResult<()> {
+        let inner = self.inner.write().await;
+        let txn = inner.db.begin().await?;
+        for fragment_id in fragment_ids {
+            let fragment = Fragment::find_by_id(fragment_id)
+                .one(&txn)
+                .await?
+                .context(format!("fragment {} not found", fragment_id))?;
+            let mut node = fragment.stream_node.to_protobuf();
+            if crate::stream::fill_snapshot_backfill_epoch(&mut node, upstream_mv_snapshot_epoch)? {
+                let node = StreamNode::from(&node);
+                Fragment::update(fragment::ActiveModel {
+                    fragment_id: Set(fragment_id),
+                    stream_node: Set(node),
+                    ..Default::default()
+                })
+                .exec(&txn)
+                .await?;
+            }
+        }
+        txn.commit().await?;
+        Ok(())
+    }
+
     /// Get the actor ids of the fragment with `fragment_id` with `Running` status.
     pub async fn get_running_actors_of_fragment(
         &self,
@@ -1639,11 +1667,11 @@ mod tests {
         let mut input = vec![];
         for (upstream_fragment_id, upstream_actor_ids) in actor_upstream_actor_ids {
             input.push(PbStreamNode {
-                node_body: Some(PbNodeBody::Merge(MergeNode {
+                node_body: Some(PbNodeBody::Merge(Box::new(MergeNode {
                     upstream_actor_id: upstream_actor_ids.iter().map(|id| *id as _).collect(),
                     upstream_fragment_id: *upstream_fragment_id as _,
                     ..Default::default()
-                })),
+                }))),
                 ..Default::default()
             });
         }
diff --git a/src/meta/src/dashboard/mod.rs b/src/meta/src/dashboard/mod.rs
index ed52378c700fc..22ebf74acb47e 100644
--- a/src/meta/src/dashboard/mod.rs
+++ b/src/meta/src/dashboard/mod.rs
@@ -49,6 +49,7 @@ pub struct DashboardService {
 pub type Service = Arc<DashboardService>;
 
 pub(super) mod handlers {
+    use std::cmp::min;
     use std::collections::HashMap;
 
     use anyhow::Context;
@@ -488,7 +489,28 @@ pub(super) mod handlers {
             let result = result
                 .map_err(|_| anyhow!("Failed to get back pressure"))
                 .map_err(err)?;
+            // TODO(eric): aggregate back_pressure_infos here
             all.back_pressure_infos.extend(result.back_pressure_infos);
+
+            // Aggregate fragment_stats
+            for (fragment_id, fragment_stats) in result.fragment_stats {
+                if let Some(s) = all.fragment_stats.get_mut(&fragment_id) {
+                    s.actor_count += fragment_stats.actor_count;
+                    s.current_epoch = min(s.current_epoch, fragment_stats.current_epoch);
+                } else {
+                    all.fragment_stats.insert(fragment_id, fragment_stats);
+                }
+            }
+
+            // Aggregate relation_stats
+            for (relation_id, relation_stats) in result.relation_stats {
+                if let Some(s) = all.relation_stats.get_mut(&relation_id) {
+                    s.actor_count += relation_stats.actor_count;
+                    s.current_epoch = min(s.current_epoch, relation_stats.current_epoch);
+                } else {
+                    all.relation_stats.insert(relation_id, relation_stats);
+                }
+            }
         }
 
         Ok(all.into())
diff --git a/src/meta/src/hummock/manager/time_travel.rs b/src/meta/src/hummock/manager/time_travel.rs
index 3e333ecafb94c..acd851b3b2781 100644
--- a/src/meta/src/hummock/manager/time_travel.rs
+++ b/src/meta/src/hummock/manager/time_travel.rs
@@ -484,6 +484,8 @@ impl HummockManager {
                 .on_conflict_do_nothing()
                 .exec(txn)
                 .await?;
+            // Return early to skip persisting delta.
+            return Ok(version_sst_ids);
         }
         let written = write_sstable_infos(
             delta.newly_added_sst_infos().filter(|s| {
diff --git a/src/meta/src/rpc/ddl_controller.rs b/src/meta/src/rpc/ddl_controller.rs
index b3c8b47b29af2..89200b4b63d6c 100644
--- a/src/meta/src/rpc/ddl_controller.rs
+++ b/src/meta/src/rpc/ddl_controller.rs
@@ -871,7 +871,7 @@ impl DdlController {
                                             format!("ProjectExecutor(from sink {})", sink_id);
                                     }
 
-                                    *merge_node = MergeNode {
+                                    **merge_node = MergeNode {
                                         upstream_actor_id: sink_actor_ids.clone(),
                                         upstream_fragment_id,
                                         upstream_dispatcher_type: DispatcherType::Hash as _,
diff --git a/src/meta/src/stream/stream_graph.rs b/src/meta/src/stream/stream_graph.rs
index 3268dcb4d16c1..c1e7763e94b88 100644
--- a/src/meta/src/stream/stream_graph.rs
+++ b/src/meta/src/stream/stream_graph.rs
@@ -18,5 +18,7 @@ mod id;
 mod schedule;
 
 pub use actor::{ActorGraphBuildResult, ActorGraphBuilder};
-pub use fragment::{CompleteStreamFragmentGraph, StreamFragmentGraph};
+pub use fragment::{
+    fill_snapshot_backfill_epoch, CompleteStreamFragmentGraph, StreamFragmentGraph,
+};
 pub use schedule::Locations;
diff --git a/src/meta/src/stream/stream_graph/actor.rs b/src/meta/src/stream/stream_graph/actor.rs
index 093f45b27958d..3d290605a38a9 100644
--- a/src/meta/src/stream/stream_graph/actor.rs
+++ b/src/meta/src/stream/stream_graph/actor.rs
@@ -151,12 +151,12 @@ impl ActorBuilder {
                 }];
 
                 Ok(StreamNode {
-                    node_body: Some(NodeBody::Merge(MergeNode {
+                    node_body: Some(NodeBody::Merge(Box::new(MergeNode {
                         upstream_actor_id: upstreams.actors.as_global_ids(),
                         upstream_fragment_id: upstreams.fragment_id.as_global_id(),
                         upstream_dispatcher_type: exchange.get_strategy()?.r#type,
                         fields: stream_node.get_fields().clone(),
-                    })),
+                    }))),
                     identity: "MergeExecutor".to_owned(),
                     ..stream_node.clone()
                 })
@@ -197,12 +197,12 @@ impl ActorBuilder {
                 let input = vec![
                     // Fill the merge node body with correct upstream info.
                     StreamNode {
-                        node_body: Some(NodeBody::Merge(MergeNode {
+                        node_body: Some(NodeBody::Merge(Box::new(MergeNode {
                             upstream_actor_id,
                             upstream_fragment_id: upstreams.fragment_id.as_global_id(),
                             upstream_dispatcher_type,
                             fields: merge_node.fields.clone(),
-                        })),
+                        }))),
                         ..merge_node.clone()
                     },
                     batch_plan_node.clone(),
@@ -246,12 +246,12 @@ impl ActorBuilder {
                 let input = vec![
                     // Fill the merge node body with correct upstream info.
                     StreamNode {
-                        node_body: Some(NodeBody::Merge(MergeNode {
+                        node_body: Some(NodeBody::Merge(Box::new(MergeNode {
                             upstream_actor_id,
                             upstream_fragment_id: upstreams.fragment_id.as_global_id(),
                             upstream_dispatcher_type: DispatcherType::NoShuffle as _,
                             fields: merge_node.fields.clone(),
-                        })),
+                        }))),
                         ..merge_node.clone()
                     },
                 ];
diff --git a/src/meta/src/stream/stream_graph/fragment.rs b/src/meta/src/stream/stream_graph/fragment.rs
index 21103fe1460c1..1203dbfe93d0a 100644
--- a/src/meta/src/stream/stream_graph/fragment.rs
+++ b/src/meta/src/stream/stream_graph/fragment.rs
@@ -27,7 +27,9 @@ use risingwave_common::catalog::{
 use risingwave_common::hash::VnodeCount;
 use risingwave_common::util::iter_util::ZipEqFast;
 use risingwave_common::util::stream_graph_visitor;
-use risingwave_common::util::stream_graph_visitor::visit_stream_node_cont;
+use risingwave_common::util::stream_graph_visitor::{
+    visit_stream_node_cont, visit_stream_node_cont_mut,
+};
 use risingwave_meta_model::WorkerId;
 use risingwave_pb::catalog::Table;
 use risingwave_pb::ddl_service::TableJobType;
@@ -38,7 +40,7 @@ use risingwave_pb::stream_plan::stream_fragment_graph::{
 use risingwave_pb::stream_plan::stream_node::NodeBody;
 use risingwave_pb::stream_plan::{
     DispatchStrategy, DispatcherType, FragmentTypeFlag, StreamActor,
-    StreamFragmentGraph as StreamFragmentGraphProto, StreamScanNode, StreamScanType,
+    StreamFragmentGraph as StreamFragmentGraphProto, StreamNode, StreamScanNode, StreamScanType,
 };
 
 use crate::barrier::SnapshotBackfillInfo;
@@ -602,8 +604,8 @@ impl StreamFragmentGraph {
                             match (prev_snapshot_backfill_info, is_snapshot_backfill) {
                                 (Some(prev_snapshot_backfill_info), true) => {
                                     prev_snapshot_backfill_info
-                                        .upstream_mv_table_ids
-                                        .insert(TableId::new(stream_scan.table_id));
+                                        .upstream_mv_table_id_to_backfill_epoch
+                                        .insert(TableId::new(stream_scan.table_id), None);
                                     true
                                 }
                                 (None, false) => true,
@@ -617,14 +619,14 @@ impl StreamFragmentGraph {
                             prev_stream_scan = Some((
                                 if is_snapshot_backfill {
                                     Some(SnapshotBackfillInfo {
-                                        upstream_mv_table_ids: HashSet::from_iter([TableId::new(
-                                            stream_scan.table_id,
-                                        )]),
+                                        upstream_mv_table_id_to_backfill_epoch: HashMap::from_iter(
+                                            [(TableId::new(stream_scan.table_id), None)],
+                                        ),
                                     })
                                 } else {
                                     None
                                 },
-                                stream_scan.clone(),
+                                *stream_scan.clone(),
                             ));
                             true
                         }
@@ -642,6 +644,44 @@ impl StreamFragmentGraph {
     }
 }
 
+/// Fill snapshot epoch for `StreamScanNode` of `SnapshotBackfill`.
+/// Return `true` when has change applied.
+pub fn fill_snapshot_backfill_epoch(
+    node: &mut StreamNode,
+    upstream_mv_table_snapshot_epoch: &HashMap<TableId, Option<u64>>,
+) -> MetaResult<bool> {
+    let mut result = Ok(());
+    let mut applied = false;
+    visit_stream_node_cont_mut(node, |node| {
+        if let Some(NodeBody::StreamScan(stream_scan)) = node.node_body.as_mut()
+            && stream_scan.stream_scan_type == StreamScanType::SnapshotBackfill as i32
+        {
+            result = try {
+                let table_id = TableId::new(stream_scan.table_id);
+                let snapshot_epoch = upstream_mv_table_snapshot_epoch
+                    .get(&table_id)
+                    .ok_or_else(|| anyhow!("upstream table id not covered: {}", table_id))?
+                    .ok_or_else(|| anyhow!("upstream table id not set: {}", table_id))?;
+                if let Some(prev_snapshot_epoch) =
+                    stream_scan.snapshot_backfill_epoch.replace(snapshot_epoch)
+                {
+                    Err(anyhow!(
+                        "snapshot backfill epoch set again: {} {} {}",
+                        table_id,
+                        prev_snapshot_epoch,
+                        snapshot_epoch
+                    ))?;
+                }
+                applied = true;
+            };
+            result.is_ok()
+        } else {
+            true
+        }
+    });
+    result.map(|_| applied)
+}
+
 static EMPTY_HASHMAP: LazyLock<HashMap<GlobalFragmentId, StreamFragmentEdge>> =
     LazyLock::new(HashMap::new);
 
diff --git a/src/meta/src/stream/test_fragmenter.rs b/src/meta/src/stream/test_fragmenter.rs
index ce844a6c764cb..3d64f378fcf5f 100644
--- a/src/meta/src/stream/test_fragmenter.rs
+++ b/src/meta/src/stream/test_fragmenter.rs
@@ -206,14 +206,14 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
         })
         .collect_vec();
     let source_node = StreamNode {
-        node_body: Some(NodeBody::Source(SourceNode {
+        node_body: Some(NodeBody::Source(Box::new(SourceNode {
             source_inner: Some(StreamSource {
                 source_id: 1,
                 state_table: Some(make_source_internal_table(0)),
                 columns,
                 ..Default::default()
             }),
-        })),
+        }))),
         stream_key: vec![2],
         ..Default::default()
     };
@@ -228,13 +228,13 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
 
     // exchange node
     let exchange_node = StreamNode {
-        node_body: Some(NodeBody::Exchange(ExchangeNode {
+        node_body: Some(NodeBody::Exchange(Box::new(ExchangeNode {
             strategy: Some(DispatchStrategy {
                 r#type: DispatcherType::Hash as i32,
                 dist_key_indices: vec![0],
                 output_indices: vec![0, 1, 2],
             }),
-        })),
+        }))),
         fields: vec![
             make_field(TypeName::Int32),
             make_field(TypeName::Int32),
@@ -252,7 +252,7 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
         children: vec![make_inputref(0), make_inputref(1)],
     };
     let filter_node = StreamNode {
-        node_body: Some(NodeBody::Filter(FilterNode {
+        node_body: Some(NodeBody::Filter(Box::new(FilterNode {
             search_condition: Some(ExprNode {
                 function_type: GreaterThan as i32,
                 return_type: Some(DataType {
@@ -261,7 +261,7 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
                 }),
                 rex_node: Some(RexNode::FuncCall(function_call)),
             }),
-        })),
+        }))),
         fields: vec![], // TODO: fill this later
         input: vec![exchange_node],
         stream_key: vec![0, 1],
@@ -272,14 +272,14 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
 
     // simple agg node
     let simple_agg_node = StreamNode {
-        node_body: Some(NodeBody::SimpleAgg(SimpleAggNode {
+        node_body: Some(NodeBody::SimpleAgg(Box::new(SimpleAggNode {
             agg_calls: vec![make_sum_aggcall(0), make_sum_aggcall(1)],
             distribution_key: Default::default(),
             is_append_only: false,
             agg_call_states: vec![make_agg_call_result_state(), make_agg_call_result_state()],
             intermediate_state_table: Some(make_empty_table(1)),
             ..Default::default()
-        })),
+        }))),
         input: vec![filter_node],
         fields: vec![], // TODO: fill this later
         stream_key: vec![0, 1],
@@ -299,12 +299,12 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
 
     // exchange node
     let exchange_node_1 = StreamNode {
-        node_body: Some(NodeBody::Exchange(ExchangeNode {
+        node_body: Some(NodeBody::Exchange(Box::new(ExchangeNode {
             strategy: Some(DispatchStrategy {
                 r#type: DispatcherType::Simple as i32,
                 ..Default::default()
             }),
-        })),
+        }))),
         fields: vec![make_field(TypeName::Int64), make_field(TypeName::Int64)],
         input: vec![],
         stream_key: vec![0, 1],
@@ -315,14 +315,14 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
 
     // agg node
     let simple_agg_node_1 = StreamNode {
-        node_body: Some(NodeBody::SimpleAgg(SimpleAggNode {
+        node_body: Some(NodeBody::SimpleAgg(Box::new(SimpleAggNode {
             agg_calls: vec![make_sum_aggcall(0), make_sum_aggcall(1)],
             distribution_key: Default::default(),
             is_append_only: false,
             agg_call_states: vec![make_agg_call_result_state(), make_agg_call_result_state()],
             intermediate_state_table: Some(make_empty_table(2)),
             ..Default::default()
-        })),
+        }))),
         fields: vec![], // TODO: fill this later
         input: vec![exchange_node_1],
         stream_key: vec![0, 1],
@@ -336,7 +336,7 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
         children: vec![make_inputref(0), make_inputref(1)],
     };
     let project_node = StreamNode {
-        node_body: Some(NodeBody::Project(ProjectNode {
+        node_body: Some(NodeBody::Project(Box::new(ProjectNode {
             select_list: vec![
                 ExprNode {
                     rex_node: Some(RexNode::FuncCall(function_call_1)),
@@ -350,7 +350,7 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
                 make_inputref(1),
             ],
             ..Default::default()
-        })),
+        }))),
         fields: vec![], // TODO: fill this later
         input: vec![simple_agg_node_1],
         stream_key: vec![1, 2],
@@ -363,11 +363,11 @@ fn make_stream_fragments() -> Vec<StreamFragment> {
     let mview_node = StreamNode {
         input: vec![project_node],
         stream_key: vec![],
-        node_body: Some(NodeBody::Materialize(MaterializeNode {
+        node_body: Some(NodeBody::Materialize(Box::new(MaterializeNode {
             table_id: 1,
             table: Some(make_materialize_table(888)),
             column_orders: vec![make_column_order(1), make_column_order(2)],
-        })),
+        }))),
         fields: vec![], // TODO: fill this later
         operator_id: 7,
         identity: "MaterializeExecutor".to_owned(),
diff --git a/src/object_store/Cargo.toml b/src/object_store/Cargo.toml
index 54bc6b4524f66..0e7d213a72321 100644
--- a/src/object_store/Cargo.toml
+++ b/src/object_store/Cargo.toml
@@ -15,6 +15,7 @@ workspace = true
 async-trait = "0.1"
 await-tree = { workspace = true }
 aws-config = { workspace = true }
+# add the dependency explicitly to enable hardcoded-credentials feature
 aws-credential-types = { workspace = true }
 aws-sdk-s3 = { version = "0.5", package = "madsim-aws-sdk-s3" }
 aws-smithy-http = { workspace = true }
@@ -23,12 +24,10 @@ aws-smithy-runtime-api = { workspace = true }
 aws-smithy-types = { workspace = true }
 bytes = { version = "1", features = ["serde"] }
 crc32fast = "1"
-either = "1"
 fail = "0.5"
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
 hyper = { version = "0.14", features = ["tcp", "client"] }                    # TODO(http-bump): required by aws sdk
 hyper-rustls = { version = "0.24.2", features = ["webpki-roots"] }
-hyper-tls = "0.5.0"
 itertools = { workspace = true }
 madsim = "0.2.31"
 opendal = { workspace = true, features = [
@@ -56,11 +55,9 @@ tracing = "0.1"
 # This crate is excluded from hakari (see hakari.toml) after hdfs is introduced...## [target.'cfg(not(madsim))'.dependencies]
 # workspace-hack = { path = "../workspace-hack" }
 #
-# [package.metadata.cargo-machete]
-# ignored = ["workspace-hack"]
 #
-# [package.metadata.cargo-udeps.ignore]
-# normal = ["workspace-hack"]
+#
+#
 
 # [features]
 # hdfs-backend = ["opendal/services-hdfs", "dep:risingwave_jni_core"]
diff --git a/src/prost/Cargo.toml b/src/prost/Cargo.toml
index 230a654e0fcee..c9fcb69dafd9b 100644
--- a/src/prost/Cargo.toml
+++ b/src/prost/Cargo.toml
@@ -21,6 +21,9 @@ tonic = { workspace = true }
 [target.'cfg(not(madsim))'.dependencies]
 workspace-hack = { path = "../workspace-hack" }
 
+[dev-dependencies]
+static_assertions = "1"
+
 [build-dependencies]
 fs-err = "3.0"
 pbjson-build = "0.7"
@@ -28,11 +31,5 @@ prost-build = { workspace = true }
 tonic-build = { workspace = true }
 walkdir = "2"
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [lints]
 workspace = true
diff --git a/src/prost/build.rs b/src/prost/build.rs
index e906c47efbd55..554fc1d67b446 100644
--- a/src/prost/build.rs
+++ b/src/prost/build.rs
@@ -112,6 +112,55 @@ fn main() -> Result<(), Box<dyn std::error::Error>> {
             "#[derive(::enum_as_inner::EnumAsInner)]",
         )
         .btree_map(btree_map_paths)
+        // node body is a very large enum, so we box it to avoid stack overflow.
+        // TODO: ideally we should box all enum variants automatically https://github.com/tokio-rs/prost/issues/1209
+        .boxed(".stream_plan.StreamNode.node_body.source")
+        .boxed(".stream_plan.StreamNode.node_body.project")
+        .boxed(".stream_plan.StreamNode.node_body.filter")
+        .boxed(".stream_plan.StreamNode.node_body.materialize")
+        .boxed(".stream_plan.StreamNode.node_body.stateless_simple_agg")
+        .boxed(".stream_plan.StreamNode.node_body.simple_agg")
+        .boxed(".stream_plan.StreamNode.node_body.hash_agg")
+        .boxed(".stream_plan.StreamNode.node_body.append_only_top_n")
+        .boxed(".stream_plan.StreamNode.node_body.hash_join")
+        .boxed(".stream_plan.StreamNode.node_body.top_n")
+        .boxed(".stream_plan.StreamNode.node_body.hop_window")
+        .boxed(".stream_plan.StreamNode.node_body.merge")
+        .boxed(".stream_plan.StreamNode.node_body.exchange")
+        .boxed(".stream_plan.StreamNode.node_body.stream_scan")
+        .boxed(".stream_plan.StreamNode.node_body.batch_plan")
+        .boxed(".stream_plan.StreamNode.node_body.lookup")
+        .boxed(".stream_plan.StreamNode.node_body.arrange")
+        .boxed(".stream_plan.StreamNode.node_body.lookup_union")
+        .boxed(".stream_plan.StreamNode.node_body.delta_index_join")
+        .boxed(".stream_plan.StreamNode.node_body.sink")
+        .boxed(".stream_plan.StreamNode.node_body.expand")
+        .boxed(".stream_plan.StreamNode.node_body.dynamic_filter")
+        .boxed(".stream_plan.StreamNode.node_body.project_set")
+        .boxed(".stream_plan.StreamNode.node_body.group_top_n")
+        .boxed(".stream_plan.StreamNode.node_body.sort")
+        .boxed(".stream_plan.StreamNode.node_body.watermark_filter")
+        .boxed(".stream_plan.StreamNode.node_body.dml")
+        .boxed(".stream_plan.StreamNode.node_body.row_id_gen")
+        .boxed(".stream_plan.StreamNode.node_body.now")
+        .boxed(".stream_plan.StreamNode.node_body.append_only_group_top_n")
+        .boxed(".stream_plan.StreamNode.node_body.temporal_join")
+        .boxed(".stream_plan.StreamNode.node_body.barrier_recv")
+        .boxed(".stream_plan.StreamNode.node_body.values")
+        .boxed(".stream_plan.StreamNode.node_body.append_only_dedup")
+        .boxed(".stream_plan.StreamNode.node_body.eowc_over_window")
+        .boxed(".stream_plan.StreamNode.node_body.over_window")
+        .boxed(".stream_plan.StreamNode.node_body.stream_fs_fetch")
+        .boxed(".stream_plan.StreamNode.node_body.stream_cdc_scan")
+        .boxed(".stream_plan.StreamNode.node_body.cdc_filter")
+        .boxed(".stream_plan.StreamNode.node_body.source_backfill")
+        .boxed(".stream_plan.StreamNode.node_body.changelog")
+        .boxed(".stream_plan.StreamNode.node_body.local_approx_percentile")
+        .boxed(".stream_plan.StreamNode.node_body.global_approx_percentile")
+        .boxed(".stream_plan.StreamNode.node_body.row_merge")
+        .boxed(".stream_plan.StreamNode.node_body.as_of_join")
+        // `Udf` is 248 bytes, while 2nd largest field is 32 bytes.
+        .boxed(".expr.ExprNode.rex_node.udf")
         // Eq + Hash are for plan nodes to do common sub-plan detection.
         // The requirement is from Source node -> SourceCatalog -> WatermarkDesc -> expr
         .type_attribute("catalog.WatermarkDesc", "#[derive(Eq, Hash)]")
diff --git a/src/prost/helpers/Cargo.toml b/src/prost/helpers/Cargo.toml
index c78ac3f2a8ece..177b1b97dadaa 100644
--- a/src/prost/helpers/Cargo.toml
+++ b/src/prost/helpers/Cargo.toml
@@ -11,11 +11,5 @@ proc-macro2 = { version = "1", default-features = false }
 quote = "1"
 syn = "2"
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [lints]
 workspace = true
diff --git a/src/prost/src/lib.rs b/src/prost/src/lib.rs
index 8fa584522354d..294d5d9514bcf 100644
--- a/src/prost/src/lib.rs
+++ b/src/prost/src/lib.rs
@@ -13,8 +13,15 @@
 // limitations under the License.
 
 // for derived code of `Message`
-#![expect(clippy::all)]
 #![expect(clippy::doc_markdown)]
+#![expect(clippy::upper_case_acronyms)]
+#![expect(clippy::needless_lifetimes)]
+// For tonic::transport::Endpoint::connect
+#![expect(clippy::disallowed_methods)]
+#![expect(clippy::enum_variant_names)]
+#![expect(clippy::module_inception)]
+// FIXME: This should be fixed!!! https://github.com/risingwavelabs/risingwave/issues/19906
+#![expect(clippy::large_enum_variant)]
 
 use std::collections::HashMap;
 use std::str::FromStr;
@@ -496,11 +503,12 @@ impl std::fmt::Debug for plan_common::ColumnDesc {
             s.field("additional_column_type", additional_column_type);
         }
         s.field("version", version);
-        if let Some(AdditionalColumn { column_type }) = additional_column {
+        if let Some(AdditionalColumn {
+            column_type: Some(column_type),
+        }) = additional_column
+        {
             // AdditionalColumn { None } means a normal column
-            if let Some(column_type) = column_type {
-                s.field("additional_column", &column_type);
-            }
+            s.field("additional_column", &column_type);
         }
         if let Some(generated_or_default_column) = generated_or_default_column {
             s.field("generated_or_default_column", &generated_or_default_column);
@@ -513,11 +521,14 @@ impl std::fmt::Debug for plan_common::ColumnDesc {
 mod tests {
     use crate::data::{data_type, DataType};
     use crate::plan_common::Field;
+    use crate::stream_plan::stream_node::NodeBody;
 
     #[test]
     fn test_getter() {
-        let mut data_type: DataType = DataType::default();
-        data_type.is_nullable = true;
+        let data_type: DataType = DataType {
+            is_nullable: true,
+            ..Default::default()
+        };
         let field = Field {
             data_type: Some(data_type),
             name: "".to_owned(),
@@ -551,4 +562,12 @@ mod tests {
         };
         assert!(!new_data_type.is_nullable);
     }
+
+    #[test]
+    fn test_size() {
+        use static_assertions::const_assert_eq;
+        // box all fields in NodeBody to avoid large_enum_variant
+        // see https://github.com/risingwavelabs/risingwave/issues/19910
+        const_assert_eq!(std::mem::size_of::<NodeBody>(), 16);
+    }
 }
diff --git a/src/risedevtool/Cargo.toml b/src/risedevtool/Cargo.toml
index 2c415d9f5da78..3334ea6bd4162 100644
--- a/src/risedevtool/Cargo.toml
+++ b/src/risedevtool/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = { version = "1", features = ["backtrace"] }
 chrono = { version = "0.4", default-features = false, features = [
@@ -32,7 +26,6 @@ redis = "0.25"
 regex = "1"
 reqwest = { version = "0.12.2", features = ["blocking"] }
 serde = { version = "1", features = ["derive"] }
-serde_json = "1"
 serde_with = "3"
 serde_yaml = "0.9"
 sqlx = { workspace = true, features = ["any"] }
diff --git a/src/rpc_client/Cargo.toml b/src/rpc_client/Cargo.toml
index 1da4027c5110c..d10a7d8d766d7 100644
--- a/src/rpc_client/Cargo.toml
+++ b/src/rpc_client/Cargo.toml
@@ -7,22 +7,13 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 async-trait = "0.1"
 easy-ext = "1"
 either = "1.13.0"
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
-h2 = "0.4.6" # https://github.com/risingwavelabs/risingwave/issues/18039
 http = "1"
-hyper = "1"
-itertools = { workspace = true }
 lru = { workspace = true }
 moka = { version = "0.12.0", features = ["future"] }
 paste = "1"
@@ -48,7 +39,6 @@ tokio-stream = { workspace = true }
 tonic = { workspace = true }
 tower = "0.5"
 tracing = "0.1"
-url = "2.5.0"
 
 [dev-dependencies]
 risingwave_hummock_sdk = { workspace = true }
diff --git a/src/sqlparser/Cargo.toml b/src/sqlparser/Cargo.toml
index 7638d27208beb..18c1068122d34 100644
--- a/src/sqlparser/Cargo.toml
+++ b/src/sqlparser/Cargo.toml
@@ -15,12 +15,6 @@ path = "src/lib.rs"
 default = ["std"]
 std = []
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 itertools = { workspace = true }
 serde = { version = "1.0", features = ["derive"], optional = true }
diff --git a/src/sqlparser/fuzz/Cargo.toml b/src/sqlparser/fuzz/Cargo.toml
index 24ebb8e6ba7fd..c031295f78fec 100644
--- a/src/sqlparser/fuzz/Cargo.toml
+++ b/src/sqlparser/fuzz/Cargo.toml
@@ -4,11 +4,9 @@ version = "0.1.0"
 edition = "2018"
 publish = false
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
 
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
+
+
 
 [dependencies]
 honggfuzz = "0.5.54"
diff --git a/src/sqlparser/sqlparser_bench/Cargo.toml b/src/sqlparser/sqlparser_bench/Cargo.toml
index f28d7ef75e2a2..a654ad5a6fb94 100644
--- a/src/sqlparser/sqlparser_bench/Cargo.toml
+++ b/src/sqlparser/sqlparser_bench/Cargo.toml
@@ -4,11 +4,9 @@ version = "0.1.0"
 authors = ["Dandandan <danielheres@gmail.com>"]
 edition = "2018"
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
 
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
+
+
 
 [dependencies]
 risingwave_sqlparser = { workspace = true }
diff --git a/src/sqlparser/src/ast/mod.rs b/src/sqlparser/src/ast/mod.rs
index c0fce4a2d4780..ec525ddcc5951 100644
--- a/src/sqlparser/src/ast/mod.rs
+++ b/src/sqlparser/src/ast/mod.rs
@@ -2839,7 +2839,7 @@ impl fmt::Display for EmitMode {
     }
 }
 
-#[derive(Debug, Clone, PartialEq, Eq, Hash)]
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 pub enum OnConflict {
     UpdateFull,
diff --git a/src/storage/Cargo.toml b/src/storage/Cargo.toml
index a5ccd017bdd75..39dc955964105 100644
--- a/src/storage/Cargo.toml
+++ b/src/storage/Cargo.toml
@@ -7,22 +7,13 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
-ahash = "0.8.7"
 anyhow = "1"
 arc-swap = "1"
 async-trait = "0.1"
 auto_enums = { workspace = true }
 await-tree = { workspace = true }
 bytes = { version = "1", features = ["serde"] }
-crossbeam = "0.8.2"
-dashmap = { version = "6", default-features = false }
 dyn-clone = "1.0.14"
 either = "1"
 enum-as-inner = "0.6"
diff --git a/src/storage/backup/Cargo.toml b/src/storage/backup/Cargo.toml
index 23a2c99567c48..f9b88044eb69e 100644
--- a/src/storage/backup/Cargo.toml
+++ b/src/storage/backup/Cargo.toml
@@ -7,16 +7,9 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 async-trait = "0.1"
-bincode = "1.3"
 bytes = { version = "1", features = ["serde"] }
 itertools = { workspace = true }
 parking_lot = { workspace = true }
diff --git a/src/storage/compactor/Cargo.toml b/src/storage/compactor/Cargo.toml
index ed8f421ee6384..d94da125ca3ec 100644
--- a/src/storage/compactor/Cargo.toml
+++ b/src/storage/compactor/Cargo.toml
@@ -8,18 +8,11 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 async-trait = "0.1"
 await-tree = { workspace = true }
 clap = { workspace = true }
 jsonbb = { workspace = true }
-parking_lot = { workspace = true }
 prost = { workspace = true }
 risingwave_common = { workspace = true }
 risingwave_common_heap_profiling = { workspace = true }
diff --git a/src/storage/hummock_sdk/Cargo.toml b/src/storage/hummock_sdk/Cargo.toml
index c6dff5afb32d9..8c911f12b5a91 100644
--- a/src/storage/hummock_sdk/Cargo.toml
+++ b/src/storage/hummock_sdk/Cargo.toml
@@ -7,25 +7,16 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack", "num-traits"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 bincode = { version = "=2.0.0-rc.3", features = ["serde"] }
 bytes = "1"
-easy-ext = "1"
 hex = "0.4"
 itertools = { workspace = true }
 parse-display = "0.10"
-prost = { workspace = true }
 risingwave_common = { workspace = true }
 risingwave_common_estimate_size = { workspace = true }
 risingwave_pb = { workspace = true }
 serde = { version = "1", features = ["derive"] }
-serde_bytes = "0.11"
 tracing = "0.1"
 
 [target.'cfg(not(madsim))'.dependencies]
diff --git a/src/storage/hummock_sdk/src/version.rs b/src/storage/hummock_sdk/src/version.rs
index 7f4b4dafb1bf7..f0bd4582752be 100644
--- a/src/storage/hummock_sdk/src/version.rs
+++ b/src/storage/hummock_sdk/src/version.rs
@@ -271,6 +271,21 @@ impl HummockVersion {
                 .values()
                 .map(|table_watermark| table_watermark.estimated_encode_len())
                 .sum::<usize>()
+            + self
+                .table_change_log
+                .values()
+                .map(|c| {
+                    c.0.iter()
+                        .map(|l| {
+                            l.old_value
+                                .iter()
+                                .chain(l.new_value.iter())
+                                .map(|s| s.estimated_encode_len())
+                                .sum::<usize>()
+                        })
+                        .sum::<usize>()
+                })
+                .sum::<usize>()
     }
 }
 
diff --git a/src/storage/hummock_test/Cargo.toml b/src/storage/hummock_test/Cargo.toml
index da96b5883462a..45a5ce62c5ba1 100644
--- a/src/storage/hummock_test/Cargo.toml
+++ b/src/storage/hummock_test/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 async-trait = "0.1"
 bytes = { version = "1" }
@@ -34,7 +28,7 @@ risingwave_object_store = { workspace = true }
 risingwave_pb = { workspace = true }
 risingwave_rpc_client = { workspace = true }
 risingwave_storage = { workspace = true, features = ["test"] }
-serde = { version = "1", features = ["derive"] }
+serial_test = { version = "3.2", optional = true }
 tokio = { version = "0.2", package = "madsim-tokio" }
 
 [target.'cfg(not(madsim))'.dependencies]
@@ -51,12 +45,11 @@ futures = { version = "0.3", default-features = false, features = [
 futures-async-stream = "0.2.9"
 risingwave_hummock_sdk = { workspace = true }
 risingwave_test_runner = { workspace = true }
-serial_test = "3.2"
 sync-point = { path = "../../utils/sync-point" }
 
 [features]
 failpoints = ["risingwave_storage/failpoints"]
-sync_point = ["sync-point/sync_point"]
+sync_point = ["sync-point/sync_point", "serial_test"]
 test = []
 
 [[bench]]
diff --git a/src/storage/src/hummock/event_handler/hummock_event_handler.rs b/src/storage/src/hummock/event_handler/hummock_event_handler.rs
index a2da11bee7159..912affeeae200 100644
--- a/src/storage/src/hummock/event_handler/hummock_event_handler.rs
+++ b/src/storage/src/hummock/event_handler/hummock_event_handler.rs
@@ -24,9 +24,9 @@ use await_tree::InstrumentAwait;
 use futures::FutureExt;
 use itertools::Itertools;
 use parking_lot::RwLock;
-use prometheus::core::{AtomicU64, GenericGauge};
 use prometheus::{Histogram, IntGauge};
 use risingwave_common::catalog::TableId;
+use risingwave_common::metrics::UintGauge;
 use risingwave_hummock_sdk::compaction_group::hummock_version_ext::SstDeltaInfo;
 use risingwave_hummock_sdk::{HummockEpoch, SyncResult};
 use tokio::spawn;
@@ -62,14 +62,11 @@ pub struct BufferTracker {
     flush_threshold: usize,
     min_batch_flush_size: usize,
     global_buffer: Arc<MemoryLimiter>,
-    global_upload_task_size: GenericGauge<AtomicU64>,
+    global_upload_task_size: UintGauge,
 }
 
 impl BufferTracker {
-    pub fn from_storage_opts(
-        config: &StorageOpts,
-        global_upload_task_size: GenericGauge<AtomicU64>,
-    ) -> Self {
+    pub fn from_storage_opts(config: &StorageOpts, global_upload_task_size: UintGauge) -> Self {
         let capacity = config.shared_buffer_capacity_mb * (1 << 20);
         let flush_threshold = (capacity as f32 * config.shared_buffer_flush_ratio) as usize;
         let shared_buffer_min_batch_flush_size =
@@ -93,7 +90,7 @@ impl BufferTracker {
         Self::new(
             usize::MAX,
             flush_threshold,
-            GenericGauge::new("test", "test").unwrap(),
+            UintGauge::new("test", "test").unwrap(),
             min_batch_flush_size,
         )
     }
@@ -101,7 +98,7 @@ impl BufferTracker {
     fn new(
         capacity: usize,
         flush_threshold: usize,
-        global_upload_task_size: GenericGauge<AtomicU64>,
+        global_upload_task_size: UintGauge,
         min_batch_flush_size: usize,
     ) -> Self {
         assert!(capacity >= flush_threshold);
@@ -116,7 +113,7 @@ impl BufferTracker {
     pub fn for_test() -> Self {
         Self::from_storage_opts(
             &StorageOpts::default(),
-            GenericGauge::new("test", "test").unwrap(),
+            UintGauge::new("test", "test").unwrap(),
         )
     }
 
@@ -128,7 +125,7 @@ impl BufferTracker {
         &self.global_buffer
     }
 
-    pub fn global_upload_task_size(&self) -> &GenericGauge<AtomicU64> {
+    pub fn global_upload_task_size(&self) -> &UintGauge {
         &self.global_upload_task_size
     }
 
diff --git a/src/storage/src/hummock/event_handler/uploader/mod.rs b/src/storage/src/hummock/event_handler/uploader/mod.rs
index 26a8fc436383f..4499aaa9700ea 100644
--- a/src/storage/src/hummock/event_handler/uploader/mod.rs
+++ b/src/storage/src/hummock/event_handler/uploader/mod.rs
@@ -28,10 +28,10 @@ use std::task::{ready, Context, Poll};
 use futures::FutureExt;
 use itertools::Itertools;
 use more_asserts::assert_gt;
-use prometheus::core::{AtomicU64, GenericGauge};
 use prometheus::{HistogramTimer, IntGauge};
 use risingwave_common::bitmap::BitmapBuilder;
 use risingwave_common::catalog::TableId;
+use risingwave_common::metrics::UintGauge;
 use risingwave_common::must_match;
 use risingwave_hummock_sdk::table_watermark::{
     TableWatermarks, VnodeWatermark, WatermarkDirection, WatermarkSerdeType,
@@ -112,14 +112,14 @@ mod uploader_imm {
     use std::fmt::Formatter;
     use std::ops::Deref;
 
-    use prometheus::core::{AtomicU64, GenericGauge};
+    use risingwave_common::metrics::UintGauge;
 
     use crate::hummock::event_handler::uploader::UploaderContext;
     use crate::mem_table::ImmutableMemtable;
 
     pub(super) struct UploaderImm {
         inner: ImmutableMemtable,
-        size_guard: GenericGauge<AtomicU64>,
+        size_guard: UintGauge,
     }
 
     impl UploaderImm {
@@ -137,7 +137,7 @@ mod uploader_imm {
         pub(super) fn for_test(imm: ImmutableMemtable) -> Self {
             Self {
                 inner: imm,
-                size_guard: GenericGauge::new("test", "test").unwrap(),
+                size_guard: UintGauge::new("test", "test").unwrap(),
             }
         }
     }
@@ -175,7 +175,7 @@ struct UploadingTask {
     join_handle: JoinHandle<HummockResult<UploadTaskOutput>>,
     task_info: UploadTaskInfo,
     spawn_upload_task: SpawnUploadTask,
-    task_size_guard: GenericGauge<AtomicU64>,
+    task_size_guard: UintGauge,
     task_count_guard: IntGauge,
 }
 
@@ -481,6 +481,7 @@ struct LocalInstanceUnsyncData {
     sealed_data: VecDeque<LocalInstanceEpochData>,
     // newer data comes first
     flushing_imms: VecDeque<SharedBufferBatchId>,
+    is_destroyed: bool,
 }
 
 impl LocalInstanceUnsyncData {
@@ -491,10 +492,12 @@ impl LocalInstanceUnsyncData {
             current_epoch_data: Some(LocalInstanceEpochData::new(init_epoch)),
             sealed_data: VecDeque::new(),
             flushing_imms: Default::default(),
+            is_destroyed: false,
         }
     }
 
     fn add_imm(&mut self, imm: UploaderImm) {
+        assert!(!self.is_destroyed);
         assert_eq!(self.table_id, imm.table_id);
         self.current_epoch_data
             .as_mut()
@@ -503,6 +506,7 @@ impl LocalInstanceUnsyncData {
     }
 
     fn local_seal_epoch(&mut self, next_epoch: HummockEpoch) -> HummockEpoch {
+        assert!(!self.is_destroyed);
         let data = self
             .current_epoch_data
             .as_mut()
@@ -626,6 +630,10 @@ impl LocalInstanceUnsyncData {
             }
         }
     }
+
+    fn is_finished(&self) -> bool {
+        self.is_destroyed && self.sealed_data.is_empty()
+    }
 }
 
 struct TableUnsyncData {
@@ -903,18 +911,19 @@ impl UnsyncData {
         }
     }
 
-    fn may_destroy_instance(&mut self, instance_id: LocalInstanceId) -> Option<TableUnsyncData> {
-        if let Some(table_id) = self.instance_table_id.remove(&instance_id) {
+    fn may_destroy_instance(&mut self, instance_id: LocalInstanceId) {
+        if let Some(table_id) = self.instance_table_id.get(&instance_id) {
             debug!(instance_id, "destroy instance");
-            let table_data = self.table_data.get_mut(&table_id).expect("should exist");
-            assert!(table_data.instance_data.remove(&instance_id).is_some());
-            if table_data.is_empty() {
-                Some(self.table_data.remove(&table_id).expect("should exist"))
-            } else {
-                None
-            }
-        } else {
-            None
+            let table_data = self.table_data.get_mut(table_id).expect("should exist");
+            let instance_data = table_data
+                .instance_data
+                .get_mut(&instance_id)
+                .expect("should exist");
+            assert!(
+                !instance_data.is_destroyed,
+                "cannot destroy an instance for twice"
+            );
+            instance_data.is_destroyed = true;
         }
     }
 
@@ -932,9 +941,20 @@ impl UnsyncData {
                     }
                 }
                 assert!(
-                    table_unsync_data.instance_data.is_empty(),
+                    table_unsync_data
+                        .instance_data
+                        .values()
+                        .all(|instance| instance.is_destroyed),
                     "should be clear when dropping the read version instance"
                 );
+                for instance_id in table_unsync_data.instance_data.keys() {
+                    assert_eq!(
+                        *table_id,
+                        self.instance_table_id
+                            .remove(instance_id)
+                            .expect("should exist")
+                    );
+                }
             }
         }
         debug_assert!(self
@@ -1011,6 +1031,18 @@ impl UploaderData {
                             flush_payload.insert(instance_id, payload);
                         }
                     }
+                    table_data.instance_data.retain(|instance_id, data| {
+                        // remove the finished instances
+                        if data.is_finished() {
+                            assert_eq!(
+                                self.unsync_data.instance_table_id.remove(instance_id),
+                                Some(*table_id)
+                            );
+                            false
+                        } else {
+                            true
+                        }
+                    });
                     if let Some((direction, watermarks, watermark_type)) = table_watermarks {
                         Self::add_table_watermarks(
                             &mut all_table_watermarks,
@@ -1429,28 +1461,7 @@ impl HummockUploader {
         let UploaderState::Working(data) = &mut self.state else {
             return;
         };
-        if let Some(removed_table_data) = data.unsync_data.may_destroy_instance(instance_id) {
-            data.task_manager.remove_table_spill_tasks(
-                removed_table_data.table_id,
-                removed_table_data
-                    .spill_tasks
-                    .into_values()
-                    .flat_map(|task_ids| task_ids.into_iter())
-                    .filter(|task_id| {
-                        if let Some((_, table_ids)) = data.unsync_data.spilled_data.get_mut(task_id)
-                        {
-                            assert!(table_ids.remove(&removed_table_data.table_id));
-                            if table_ids.is_empty() {
-                                data.unsync_data.spilled_data.remove(task_id);
-                            }
-                            false
-                        } else {
-                            true
-                        }
-                    }),
-            )
-        }
-        data.check_upload_task_consistency();
+        data.unsync_data.may_destroy_instance(instance_id);
     }
 
     pub(crate) fn min_uncommitted_sst_id(&self) -> Option<HummockSstableObjectId> {
@@ -1800,6 +1811,65 @@ pub(crate) mod tests {
         );
     }
 
+    #[tokio::test]
+    async fn test_uploader_destroy_instance_before_sync() {
+        let mut uploader = test_uploader(dummy_success_upload_future);
+        let epoch1 = INITIAL_EPOCH.next_epoch();
+        uploader.start_epochs_for_test([epoch1]);
+        let imm = gen_imm(epoch1).await;
+        uploader.init_instance(TEST_LOCAL_INSTANCE_ID, TEST_TABLE_ID, epoch1);
+        uploader.add_imm(TEST_LOCAL_INSTANCE_ID, imm.clone());
+        uploader.local_seal_epoch_for_test(TEST_LOCAL_INSTANCE_ID, epoch1);
+        uploader.may_destroy_instance(TEST_LOCAL_INSTANCE_ID);
+
+        let (sync_tx, sync_rx) = oneshot::channel();
+        uploader.start_single_epoch_sync(epoch1, sync_tx, HashSet::from_iter([TEST_TABLE_ID]));
+        assert_eq!(epoch1 as HummockEpoch, uploader.test_max_syncing_epoch());
+        assert_eq!(1, uploader.data().syncing_data.len());
+        let (_, syncing_data) = uploader.data().syncing_data.first_key_value().unwrap();
+        assert_eq!(epoch1 as HummockEpoch, syncing_data.sync_table_epochs[0].0);
+        assert!(syncing_data.uploaded.is_empty());
+        assert!(!syncing_data.remaining_uploading_tasks.is_empty());
+
+        let staging_sst = uploader.next_uploaded_sst().await;
+        assert_eq!(&vec![epoch1], staging_sst.epochs());
+        assert_eq!(
+            &HashMap::from_iter([(TEST_LOCAL_INSTANCE_ID, vec![imm.batch_id()])]),
+            staging_sst.imm_ids()
+        );
+        assert_eq!(
+            &dummy_success_upload_output().new_value_ssts,
+            staging_sst.sstable_infos()
+        );
+
+        match sync_rx.await {
+            Ok(Ok(data)) => {
+                let SyncedData {
+                    uploaded_ssts,
+                    table_watermarks,
+                } = data;
+                assert_eq!(1, uploaded_ssts.len());
+                let staging_sst = &uploaded_ssts[0];
+                assert_eq!(&vec![epoch1], staging_sst.epochs());
+                assert_eq!(
+                    &HashMap::from_iter([(TEST_LOCAL_INSTANCE_ID, vec![imm.batch_id()])]),
+                    staging_sst.imm_ids()
+                );
+                assert_eq!(
+                    &dummy_success_upload_output().new_value_ssts,
+                    staging_sst.sstable_infos()
+                );
+                assert!(table_watermarks.is_empty());
+            }
+            _ => unreachable!(),
+        };
+        assert!(!uploader
+            .data()
+            .unsync_data
+            .table_data
+            .contains_key(&TEST_TABLE_ID));
+    }
+
     #[tokio::test]
     async fn test_empty_uploader_sync() {
         let mut uploader = test_uploader(dummy_success_upload_future);
diff --git a/src/storage/src/hummock/event_handler/uploader/task_manager.rs b/src/storage/src/hummock/event_handler/uploader/task_manager.rs
index fd53fae1db322..c23aa648a7222 100644
--- a/src/storage/src/hummock/event_handler/uploader/task_manager.rs
+++ b/src/storage/src/hummock/event_handler/uploader/task_manager.rs
@@ -136,24 +136,6 @@ impl TaskManager {
         )
     }
 
-    pub(super) fn remove_table_spill_tasks(
-        &mut self,
-        table_id: TableId,
-        task_ids: impl Iterator<Item = UploadingTaskId>,
-    ) {
-        for task_id in task_ids {
-            let entry = self.tasks.get_mut(&task_id).expect("should exist");
-            let empty = must_match!(&mut entry.status, UploadingTaskStatus::Spilling(table_ids) => {
-                assert!(table_ids.remove(&table_id));
-                table_ids.is_empty()
-            });
-            if empty {
-                let task = self.tasks.remove(&task_id).expect("should exist").task;
-                task.join_handle.abort();
-            }
-        }
-    }
-
     pub(super) fn sync(
         &mut self,
         context: &UploaderContext,
diff --git a/src/storage/src/monitor/hummock_state_store_metrics.rs b/src/storage/src/monitor/hummock_state_store_metrics.rs
index ced2723cafea0..acdc960f0f36c 100644
--- a/src/storage/src/monitor/hummock_state_store_metrics.rs
+++ b/src/storage/src/monitor/hummock_state_store_metrics.rs
@@ -14,7 +14,7 @@
 
 use std::sync::{Arc, OnceLock};
 
-use prometheus::core::{AtomicU64, Collector, Desc, GenericCounter, GenericGauge};
+use prometheus::core::{AtomicU64, Collector, Desc, GenericCounter};
 use prometheus::{
     exponential_buckets, histogram_opts, proto, register_histogram_vec_with_registry,
     register_histogram_with_registry, register_int_counter_vec_with_registry,
@@ -23,7 +23,7 @@ use prometheus::{
 use risingwave_common::config::MetricLevel;
 use risingwave_common::metrics::{
     RelabeledCounterVec, RelabeledGuardedHistogramVec, RelabeledGuardedIntCounterVec,
-    RelabeledHistogramVec, RelabeledMetricVec,
+    RelabeledHistogramVec, RelabeledMetricVec, UintGauge,
 };
 use risingwave_common::monitor::GLOBAL_METRICS_REGISTRY;
 use risingwave_common::{
@@ -73,9 +73,9 @@ pub struct HummockStateStoreMetrics {
     pub spill_task_size_from_sealed: GenericCounter<AtomicU64>,
 
     // uploading task
-    pub uploader_uploading_task_size: GenericGauge<AtomicU64>,
+    pub uploader_uploading_task_size: UintGauge,
     pub uploader_uploading_task_count: IntGauge,
-    pub uploader_imm_size: GenericGauge<AtomicU64>,
+    pub uploader_imm_size: UintGauge,
     pub uploader_upload_task_latency: Histogram,
     pub uploader_syncing_epoch_count: IntGauge,
     pub uploader_wait_poll_latency: Histogram,
@@ -312,7 +312,7 @@ impl HummockStateStoreMetrics {
         )
         .unwrap();
 
-        let uploader_uploading_task_size = GenericGauge::new(
+        let uploader_uploading_task_size = UintGauge::new(
             "state_store_uploader_uploading_task_size",
             "Total size of uploader uploading tasks",
         )
@@ -328,7 +328,7 @@ impl HummockStateStoreMetrics {
         )
         .unwrap();
 
-        let uploader_imm_size = GenericGauge::new(
+        let uploader_imm_size = UintGauge::new(
             "state_store_uploader_imm_size",
             "Total size of imms tracked by uploader",
         )
diff --git a/src/stream/Cargo.toml b/src/stream/Cargo.toml
index 8a0061ab99ea9..0061174fbbd7b 100644
--- a/src/stream/Cargo.toml
+++ b/src/stream/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 arc-swap = "1"
@@ -25,6 +19,7 @@ await-tree = { workspace = true }
 bytes = "1"
 cfg-if = "1"
 delta_btree_map = { path = "../utils/delta_btree_map" }
+dhat = { version = "0.3", optional = true }
 educe = "0.6"
 either = "1"
 enum-as-inner = "0.6"
@@ -32,11 +27,7 @@ fail = "0.5"
 foyer = { workspace = true }
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
 futures-async-stream = { workspace = true }
-governor = { version = "0.6", default-features = false, features = [
-  "std",
-  "dashmap",
-  "jitter",
-] }
+governor = { workspace = true }
 hytra = "0.1.2"
 itertools = { workspace = true }
 jsonbb = { workspace = true }
@@ -65,7 +56,6 @@ serde = { version = "1.0", features = ["derive"] }
 serde_json = "1"
 smallvec = "1"
 static_assertions = "1"
-strum = "0.26"
 strum_macros = "0.26"
 thiserror = "1"
 thiserror-ext = { workspace = true }
@@ -91,7 +81,6 @@ workspace-hack = { path = "../workspace-hack" }
 [dev-dependencies]
 assert_matches = "1"
 criterion = { workspace = true, features = ["async_tokio", "async"] }
-dhat =  "0.3"
 expect-test = "1"
 risingwave_expr_impl = { workspace = true }
 risingwave_hummock_sdk = { workspace = true }
@@ -102,7 +91,7 @@ serde_yaml = "0.9"
 tracing-test = "0.2"
 
 [features]
-dhat-heap = []
+dhat-heap = ["dhat"]
 
 [[bench]]
 name = "stream_hash_agg"
diff --git a/src/stream/spill_test/Cargo.toml b/src/stream/spill_test/Cargo.toml
index fe4fddfe2b1c3..b979fdfa30b80 100644
--- a/src/stream/spill_test/Cargo.toml
+++ b/src/stream/spill_test/Cargo.toml
@@ -8,14 +8,8 @@ license = { workspace = true }
 repository = { workspace = true }
 
 [dependencies]
-async-trait = "0.1"
-bytes = { version = "1" }
-futures = { version = "0.3", default-features = false, features = ["alloc"] }
-futures-async-stream = "0.2.9"
 risingwave_common = { workspace = true }
-risingwave_hummock_sdk = { workspace = true }
 risingwave_hummock_test = { workspace = true, features = ["test"] }
-risingwave_storage = { workspace = true, features = ["test"] }
 risingwave_stream = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio" }
 
diff --git a/src/stream/src/executor/backfill/snapshot_backfill.rs b/src/stream/src/executor/backfill/snapshot_backfill.rs
index 7d33ac89d403d..6851d81ce790f 100644
--- a/src/stream/src/executor/backfill/snapshot_backfill.rs
+++ b/src/stream/src/executor/backfill/snapshot_backfill.rs
@@ -62,6 +62,8 @@ pub struct SnapshotBackfillExecutor<S: StateStore> {
 
     actor_ctx: ActorContextRef,
     metrics: Arc<StreamingMetrics>,
+
+    snapshot_epoch: Option<u64>,
 }
 
 impl<S: StateStore> SnapshotBackfillExecutor<S> {
@@ -76,6 +78,7 @@ impl<S: StateStore> SnapshotBackfillExecutor<S> {
         rate_limit: Option<usize>,
         barrier_rx: UnboundedReceiver<Barrier>,
         metrics: Arc<StreamingMetrics>,
+        snapshot_epoch: Option<u64>,
     ) -> Self {
         assert_eq!(&upstream.info.schema, upstream_table.schema());
         if let Some(rate_limit) = rate_limit {
@@ -94,17 +97,37 @@ impl<S: StateStore> SnapshotBackfillExecutor<S> {
             barrier_rx,
             actor_ctx,
             metrics,
+            snapshot_epoch,
         }
     }
 
     #[try_stream(ok = Message, error = StreamExecutorError)]
     async fn execute_inner(mut self) {
         debug!("snapshot backfill executor start");
-        let first_barrier = expect_first_barrier(&mut self.upstream).await?;
-        debug!(epoch = ?first_barrier.epoch, "get first upstream barrier");
+        let first_upstream_barrier = expect_first_barrier(&mut self.upstream).await?;
+        debug!(epoch = ?first_upstream_barrier.epoch, "get first upstream barrier");
         let first_recv_barrier = receive_next_barrier(&mut self.barrier_rx).await?;
         debug!(epoch = ?first_recv_barrier.epoch, "get first inject barrier");
-        let should_backfill = first_barrier.epoch != first_recv_barrier.epoch;
+        let should_backfill = if let Some(snapshot_epoch) = self.snapshot_epoch {
+            if first_upstream_barrier.epoch != first_recv_barrier.epoch {
+                assert_eq!(snapshot_epoch, first_upstream_barrier.epoch.prev);
+                true
+            } else {
+                false
+            }
+        } else {
+            // when snapshot epoch is not set, the StreamNode must be created previously and has finished the backfill
+            if cfg!(debug_assertions) {
+                panic!(
+                    "snapshot epoch not set. first_upstream_epoch: {:?}, first_recv_epoch: {:?}",
+                    first_upstream_barrier.epoch, first_recv_barrier.epoch
+                );
+            } else {
+                warn!(first_upstream_epoch = ?first_upstream_barrier.epoch, first_recv_epoch=?first_recv_barrier.epoch, "snapshot epoch not set");
+                assert_eq!(first_upstream_barrier.epoch, first_recv_barrier.epoch);
+                false
+            }
+        };
 
         let (mut barrier_epoch, mut need_report_finish) = {
             if should_backfill {
@@ -119,7 +142,7 @@ impl<S: StateStore> SnapshotBackfillExecutor<S> {
                 let mut upstream_buffer =
                     UpstreamBuffer::new(&mut self.upstream, consume_upstream_row_count);
 
-                let first_barrier_epoch = first_barrier.epoch;
+                let first_barrier_epoch = first_upstream_barrier.epoch;
 
                 // Phase 1: consume upstream snapshot
                 {
@@ -156,7 +179,7 @@ impl<S: StateStore> SnapshotBackfillExecutor<S> {
                     }
 
                     let recv_barrier = self.barrier_rx.recv().await.expect("should exist");
-                    assert_eq!(first_barrier.epoch, recv_barrier.epoch);
+                    assert_eq!(first_upstream_barrier.epoch, recv_barrier.epoch);
                     yield Message::Barrier(recv_barrier);
                 }
 
@@ -239,9 +262,9 @@ impl<S: StateStore> SnapshotBackfillExecutor<S> {
                     table_id = self.upstream_table.table_id().table_id,
                     "skip backfill"
                 );
-                assert_eq!(first_barrier.epoch, first_recv_barrier.epoch);
+                assert_eq!(first_upstream_barrier.epoch, first_recv_barrier.epoch);
                 yield Message::Barrier(first_recv_barrier);
-                (first_barrier.epoch, false)
+                (first_upstream_barrier.epoch, false)
             }
         };
         let mut upstream = self.upstream.into_executor(self.barrier_rx).execute();
diff --git a/src/stream/src/executor/monitor/streaming_stats.rs b/src/stream/src/executor/monitor/streaming_stats.rs
index f7da06608a722..ccefd82cfd2e9 100644
--- a/src/stream/src/executor/monitor/streaming_stats.rs
+++ b/src/stream/src/executor/monitor/streaming_stats.rs
@@ -191,7 +191,7 @@ pub struct StreamingMetrics {
     materialize_cache_hit_count: RelabeledGuardedIntCounterVec<3>,
     materialize_cache_total_count: RelabeledGuardedIntCounterVec<3>,
     materialize_input_row_count: RelabeledGuardedIntCounterVec<3>,
-    materialize_current_epoch: RelabeledGuardedIntGaugeVec<3>,
+    pub materialize_current_epoch: RelabeledGuardedIntGaugeVec<3>,
 }
 
 pub static GLOBAL_STREAMING_METRICS: OnceLock<StreamingMetrics> = OnceLock::new();
diff --git a/src/stream/src/task/stream_manager.rs b/src/stream/src/task/stream_manager.rs
index 1868670894182..eba1646259ea4 100644
--- a/src/stream/src/task/stream_manager.rs
+++ b/src/stream/src/task/stream_manager.rs
@@ -408,6 +408,7 @@ impl StreamActorManager {
             node.rate_limit.map(|x| x as _),
             barrier_rx,
             self.streaming_metrics.clone(),
+            node.snapshot_backfill_epoch,
         )
         .boxed();
 
diff --git a/src/test_runner/Cargo.toml b/src/test_runner/Cargo.toml
index 3b9819bd45dad..23ecd648aa57e 100644
--- a/src/test_runner/Cargo.toml
+++ b/src/test_runner/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 fail = "0.5"
 sync-point = { path = "../utils/sync-point" }
diff --git a/src/tests/compaction_test/Cargo.toml b/src/tests/compaction_test/Cargo.toml
index 3bc86649ea3b3..2351e750c2a95 100644
--- a/src/tests/compaction_test/Cargo.toml
+++ b/src/tests/compaction_test/Cargo.toml
@@ -8,28 +8,15 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
-async-trait = "0.1"
 bytes = "1"
 clap = { workspace = true }
 foyer = { workspace = true }
-futures = { version = "0.3", default-features = false, features = ["alloc"] }
-prometheus = { version = "0.13" }
-rand = { workspace = true }
 risingwave_common = { workspace = true }
 risingwave_compactor = { workspace = true }
 risingwave_hummock_sdk = { workspace = true }
-risingwave_hummock_test = { workspace = true }
-risingwave_meta = { workspace = true }
 risingwave_meta_node = { workspace = true }
-risingwave_object_store = { workspace = true }
 risingwave_pb = { workspace = true }
 risingwave_rpc_client = { workspace = true }
 risingwave_rt = { workspace = true }
diff --git a/src/tests/e2e_extended_mode/Cargo.toml b/src/tests/e2e_extended_mode/Cargo.toml
index 2e88f009742a7..8e4e16f2c41cd 100644
--- a/src/tests/e2e_extended_mode/Cargo.toml
+++ b/src/tests/e2e_extended_mode/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = { version = "1", features = ["backtrace"] }
 chrono = { version = "0.4", features = ['serde'] }
diff --git a/src/tests/libpq_test/Cargo.toml b/src/tests/libpq_test/Cargo.toml
index 84cefac33ef6e..8e618be7928ee 100644
--- a/src/tests/libpq_test/Cargo.toml
+++ b/src/tests/libpq_test/Cargo.toml
@@ -7,6 +7,5 @@ version = "0.1.0"
 edition = "2021"
 
 [dependencies]
-anyhow = "1"
 libpq = "3.0"
 clap = { workspace = true }
diff --git a/src/tests/regress/Cargo.toml b/src/tests/regress/Cargo.toml
index 65248877adf60..a5095d40a7f80 100644
--- a/src/tests/regress/Cargo.toml
+++ b/src/tests/regress/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = { version = "1", features = ["backtrace"] }
 clap = { workspace = true }
diff --git a/src/tests/simulation/Cargo.toml b/src/tests/simulation/Cargo.toml
index 76f2e263d7290..9fc7c348f6acd 100644
--- a/src/tests/simulation/Cargo.toml
+++ b/src/tests/simulation/Cargo.toml
@@ -5,15 +5,11 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [package.metadata.cargo-machete]
-ignored = ["serde"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["serde"]
+ignored = ["tikv-jemallocator"]
 
 [dependencies]
 anyhow = "1.0"
 async-trait = "0.1"
-aws-sdk-s3 = { version = "0.5", package = "madsim-aws-sdk-s3" }
 cfg-or-panic = "0.2"
 clap = { workspace = true }
 console = "0.15"
@@ -28,7 +24,6 @@ maplit = "1"
 paste = "1"
 pin-project = "1.1"
 pretty_assertions = "1"
-prometheus = { version = "0.13" }
 rand = { workspace = true }
 rand_chacha = { version = "0.3.1" }
 rdkafka = { workspace = true }
@@ -45,18 +40,16 @@ risingwave_hummock_sdk = { workspace = true }
 risingwave_meta_node = { workspace = true }
 risingwave_object_store = { workspace = true }
 risingwave_pb = { workspace = true }
-risingwave_rpc_client = { workspace = true }
 risingwave_sqlparser = { workspace = true }
 risingwave_sqlsmith = { workspace = true }
 serde = "1.0.188"
 serde_derive = "1.0.188"
 serde_json = "1.0.107"
-sqllogictest = "0.23.0"
+sqllogictest = "0.24.0"
 tempfile = "3"
 tikv-jemallocator = { workspace = true }
 tokio = { version = "0.2", package = "madsim-tokio" }
 tokio-postgres = "0.7"
-tokio-stream = { workspace = true }
 tracing = "0.1"
 tracing-subscriber = { version = "0.3", features = ["env-filter"] }
 uuid = "*"
diff --git a/src/tests/sqlsmith/Cargo.toml b/src/tests/sqlsmith/Cargo.toml
index c7cf9da9ce299..740749e58d572 100644
--- a/src/tests/sqlsmith/Cargo.toml
+++ b/src/tests/sqlsmith/Cargo.toml
@@ -7,12 +7,6 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 chrono = "0.4"
@@ -25,7 +19,6 @@ risingwave_common = { workspace = true }
 risingwave_expr = { workspace = true }
 risingwave_expr_impl = { workspace = true }
 risingwave_frontend = { workspace = true }
-risingwave_pb = { workspace = true }
 risingwave_sqlparser = { workspace = true }
 similar = "2.6.0"
 thiserror-ext = { workspace = true }
diff --git a/src/tests/state_cleaning_test/Cargo.toml b/src/tests/state_cleaning_test/Cargo.toml
index 6c12898343951..5a5fa90d3a85a 100644
--- a/src/tests/state_cleaning_test/Cargo.toml
+++ b/src/tests/state_cleaning_test/Cargo.toml
@@ -7,17 +7,10 @@ keywords = { workspace = true }
 license = { workspace = true }
 repository = { workspace = true }
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = "1"
 clap = { workspace = true }
 futures = { version = "0.3", default-features = false, features = ["alloc"] }
-prometheus = { version = "0.13" }
 regex = "1"
 risingwave_rt = { workspace = true }
 serde = { version = "1", features = ["derive"] }
diff --git a/src/utils/delta_btree_map/Cargo.toml b/src/utils/delta_btree_map/Cargo.toml
index 274a028489395..879f740e89425 100644
--- a/src/utils/delta_btree_map/Cargo.toml
+++ b/src/utils/delta_btree_map/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 educe = "0.6"
 enum-as-inner = "0.6"
diff --git a/src/utils/futures_util/Cargo.toml b/src/utils/futures_util/Cargo.toml
index 97bd794daaf8d..d8e763938e3e8 100644
--- a/src/utils/futures_util/Cargo.toml
+++ b/src/utils/futures_util/Cargo.toml
@@ -5,12 +5,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 futures = "0.3"
 pin-project-lite = "0.2"
diff --git a/src/utils/iter_util/Cargo.toml b/src/utils/iter_util/Cargo.toml
index d730c1cbc908e..2bfec2cedf147 100644
--- a/src/utils/iter_util/Cargo.toml
+++ b/src/utils/iter_util/Cargo.toml
@@ -5,12 +5,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 itertools = { workspace = true }
 
diff --git a/src/utils/local_stats_alloc/Cargo.toml b/src/utils/local_stats_alloc/Cargo.toml
index d80d3db38109c..b92b0b0e1359a 100644
--- a/src/utils/local_stats_alloc/Cargo.toml
+++ b/src/utils/local_stats_alloc/Cargo.toml
@@ -14,11 +14,5 @@ workspace-hack = { path = "../../workspace-hack" }
 
 [dev-dependencies]
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [lints]
 workspace = true
diff --git a/src/utils/pgwire/Cargo.toml b/src/utils/pgwire/Cargo.toml
index 8f0e55ff8446b..2275ed5a20571 100644
--- a/src/utils/pgwire/Cargo.toml
+++ b/src/utils/pgwire/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 anyhow = { version = "1.0", default-features = false }
 auto_enums = { workspace = true }
diff --git a/src/utils/resource_util/Cargo.toml b/src/utils/resource_util/Cargo.toml
index 9680c5054bc31..175052cdfbd16 100644
--- a/src/utils/resource_util/Cargo.toml
+++ b/src/utils/resource_util/Cargo.toml
@@ -5,12 +5,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 fs-err = "3"
 sysinfo = { version = "0.33", default-features = false, features = ["system"] }
diff --git a/src/utils/runtime/Cargo.toml b/src/utils/runtime/Cargo.toml
index b34cd499f9ddd..0ecadea5aeb9a 100644
--- a/src/utils/runtime/Cargo.toml
+++ b/src/utils/runtime/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 await-tree = { workspace = true }
 console = "0.15"
diff --git a/src/utils/sync-point/Cargo.toml b/src/utils/sync-point/Cargo.toml
index d228bfbe79a35..27b908bc517b1 100644
--- a/src/utils/sync-point/Cargo.toml
+++ b/src/utils/sync-point/Cargo.toml
@@ -4,12 +4,6 @@ version = "0.1.0"
 edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 futures-util = "0.3"
 spin = "0.9"
diff --git a/src/utils/variables/Cargo.toml b/src/utils/variables/Cargo.toml
index 7bcc1b2d963c0..c897c42986f2a 100644
--- a/src/utils/variables/Cargo.toml
+++ b/src/utils/variables/Cargo.toml
@@ -8,12 +8,6 @@ license = { workspace = true }
 repository = { workspace = true }
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
-[package.metadata.cargo-machete]
-ignored = ["workspace-hack"]
-
-[package.metadata.cargo-udeps.ignore]
-normal = ["workspace-hack"]
-
 [dependencies]
 chrono = { version = "0.4", default-features = false, features = [
     "clock",