Apply suggestions from code review

Co-authored-by: xixirangrang <[email protected]>

tidb-troubleshooting-map.md

@@ -242,9 +242,9 @@ Check the specific cause for busy by viewing the monitor **Grafana** -> **TiKV**
 
 - 4.3.1 TiKV RocksDB encounters `write stall`.
 
-    A TiKV instance has two RocksDB instances, one in `data/raft` to store the Raft log, and another in `data/db` to store the real data. You can check the specific cause for the stall by running `grep "Stalling" RocksDB` in the log. The RocksDB log is a file starting with `LOG`, and `LOG` is the current log. `write stall` is a performance degradation mechanism natively built in RocksDB. When `write stall` occurs in RocksDB, the system performance drops significantly. In versions before v5.2.0, TiDB tries to block all write requests by returning a `ServerIsBusy` error directly to the client when encountering `write stall`, but this could lead to a sharp decrease in QPS performance. Starting from v5.2.0, TiKV introduces a new flow control mechanism that suppresses writes by delaying write requests dynamically in the scheduling layer, which replaces the previous mechanism of returning `server is busy` to clients when `write stall` occurs. The new flow control mechanism is enabled by default, and TiKV automatically disables the `write stall` mechanism for `KvDB` and `RaftDB` (except for memtables). However, when the number of pending requests exceeds a certain threshold, the flow control mechanism still takes effect, beginning to reject some or all write requests and returning a `server is busy` error to clients. For detailed explanations and thresholds, see [Flow control configuration](/tikv-configuration-file#storageflow-control).
+    A TiKV instance has two RocksDB instances, one in `data/raft` to store the Raft log, and the other in `data/db` to store the real data. You can check the specific cause for the stall by running `grep "Stalling" RocksDB` in the log. The RocksDB log is a file starting with `LOG`, and `LOG` is the current log. `write stall` is a performance degradation mechanism natively built in RocksDB. When `write stall` occurs in RocksDB, the system performance drops significantly. In versions before v5.2.0, TiDB tries to block all write requests by returning a `ServerIsBusy` error directly to the client when encountering `write stall`, but this could lead to a sharp decrease in QPS performance. Starting from v5.2.0, TiKV introduces a new flow control mechanism that suppresses writes by delaying write requests dynamically in the scheduling layer, which replaces the previous mechanism of returning `server is busy` to clients when `write stall` occurs. The new flow control mechanism is enabled by default, and TiKV automatically disables the `write stall` mechanism for `KvDB` and `RaftDB` (except for memtables). However, when the number of pending requests exceeds a certain threshold, the flow control mechanism still takes effect, beginning to reject some or all write requests and returning a `server is busy` error to clients. For detailed explanations and thresholds, see [Flow control configuration](/tikv-configuration-file.md#storageflow-control).
 
-    - If the `server is busy` error is triggered by too many pending compaction bytes, you can alleviate this issue by increasing the values of the `soft-pending-compaction-bytes-limit` and `hard-pending-compaction-bytes-limit` parameters.
+    - If the `server is busy` error is triggered by too many pending compaction bytes, you can alleviate this issue by increasing the values of the  [`soft-pending-compaction-bytes-limit`](/tikv-configuration-file.md#soft-pending-compaction-bytes-limit) and [`hard-pending-compaction-bytes-limit`](/tikv-configuration-file.md#hard-pending-compaction-bytes-limit) parameters.
 
         - When pending compaction bytes reach the value of the `soft-pending-compaction-bytes-limit` parameter (`192GiB` by default), the flow control mechanism begins to reject some write requests (by returning `ServerIsBusy` to clients). In this case, you can increase the value of this parameter, for example, `[storage.flow-control] soft-pending-compaction-bytes-limit = "384GiB"`.