Add annotations for controlling request rate limiting (#4660)

docs/content/configuration/ingress-resources/advanced-configuration-with-annotations.md

@@ -175,6 +175,22 @@ The table below summarizes the available annotations.
 |``nginx.org/use-cluster-ip`` | N/A | Enables using the Cluster IP and port of the service instead of the default behavior of using the IP and port of the pods. When this field is enabled, the fields that configure NGINX behavior related to multiple upstream servers (like ``lb-method`` and ``next-upstream``) will have no effect, as NGINX Ingress Controller will configure NGINX with only one upstream server that will match the service Cluster IP.   | ``False`` |  |
 {{% /table %}}
 
+### Rate limiting
+
+{{% table %}}
+|Annotation | ConfigMap Key | Description | Default | Example |
+| ---| ---| ---| ---| --- |
+|``nginx.org/limit-req-rate`` | N/A | Enables request-rate-limiting for this ingress by creating a [limit_req_zone](https://nginx.org/en/docs/http/ngx_http_limit_req_module.html#limit_req_zone) and matching [limit_req](https://nginx.org/en/docs/http/ngx_http_limit_req_module.html#limit_req) for each location. All servers/locations of one ingress share the same zone. Must have unit r/s or r/m. | N/A | 200r/s |
+|``nginx.org/limit-req-key`` | N/A | The key to which the rate limit is applied. Can contain text, variables, or a combination of them. Variables must be surrounded by ${}. | ${binary_remote_addr} | ${binary_remote_addr} |
+|``nginx.org/limit-req-zone-size`` | N/A | Configures the size of the created [limit_req_zone](https://nginx.org/en/docs/http/ngx_http_limit_req_module.html#limit_req_zone). | 10m | 20m |
+|``nginx.org/limit-req-delay`` | N/A | Configures the delay-parameter of the [limit_req](https://nginx.org/en/docs/http/ngx_http_limit_req_module.html#limit_req) directive. | 0 | 100 |
+|``nginx.org/limit-req-no-delay`` | N/A | Configures the nodelay-parameter of the [limit_req](https://nginx.org/en/docs/http/ngx_http_limit_req_module.html#limit_req) directive. | false | true |
+|``nginx.org/limit-req-burst`` | N/A | Configures the burst-parameter of the [limit_req](https://nginx.org/en/docs/http/ngx_http_limit_req_module.html#limit_req) directive. | N/A | 100 |
+|``nginx.org/limit-req-dry-run`` | N/A | Enables the dry run mode. In this mode, the rate limit is not actually applied, but the number of excessive requests is accounted as usual in the shared memory zone. | false | true |
+|``nginx.org/limit-req-log-level`` | N/A | Sets the desired logging level for cases when the server refuses to process requests due to rate exceeding, or delays request processing. Allowed values are info, notice, warn or error. | error | info |
+|``nginx.org/limit-req-reject-code`` | N/A | Sets the status code to return in response to rejected requests. Must fall into the range 400..599. | 429 | 503 |
+{{% /table %}}
+
 ### Snippets and Custom Templates
 
 {{% table %}}

internal/configs/annotations.go

@@ -1,6 +1,9 @@
 package configs
 
 import (
+	"fmt"
+	"slices"
+
 	"github.com/golang/glog"
 )
 
@@ -78,6 +81,15 @@ var minionInheritanceList = map[string]bool{
 	"nginx.org/max-fails":                true,
 	"nginx.org/max-conns":                true,
 	"nginx.org/fail-timeout":             true,
+	"nginx.org/limit-req-rate":           true,
+	"nginx.org/limit-req-key":            true,
+	"nginx.org/limit-req-zone-size":      true,
+	"nginx.org/limit-req-delay":          true,
+	"nginx.org/limit-req-no-delay":       true,
+	"nginx.org/limit-req-burst":          true,
+	"nginx.org/limit-req-dry-run":        true,
+	"nginx.org/limit-req-log-level":      true,
+	"nginx.org/limit-req-reject-code":    true,
 }
 
 var validPathRegex = map[string]bool{
@@ -413,9 +425,81 @@ func parseAnnotations(ingEx *IngressEx, baseCfgParams *ConfigParams, isPlus bool
 			cfgParams.UseClusterIP = useClusterIP
 		}
 	}
+
+	for _, err := range parseRateLimitAnnotations(ingEx.Ingress.Annotations, &cfgParams, ingEx.Ingress) {
+		glog.Error(err)
+	}
+
 	return cfgParams
 }
 
+// parseRateLimitAnnotations parses rate-limiting-related annotations and places them into cfgParams. Occurring errors are collected and returned, but do not abort parsing.
+//
+//gocyclo:ignore
+func parseRateLimitAnnotations(annotations map[string]string, cfgParams *ConfigParams, context apiObject) []error {
+	errors := make([]error, 0)
+	if requestRateLimit, exists := annotations["nginx.org/limit-req-rate"]; exists {
+		if rate, err := ParseRequestRate(requestRateLimit); err != nil {
+			errors = append(errors, fmt.Errorf("Ingress %s/%s: Invalid value for nginx.org/limit-req-rate: got %s: %w", context.GetNamespace(), context.GetName(), requestRateLimit, err))
+		} else {
+			cfgParams.LimitReqRate = rate
+		}
+	}
+	if requestRateKey, exists := annotations["nginx.org/limit-req-key"]; exists {
+		cfgParams.LimitReqKey = requestRateKey
+	}
+	if requestRateZoneSize, exists := annotations["nginx.org/limit-req-zone-size"]; exists {
+		if size, err := ParseSize(requestRateZoneSize); err != nil {
+			errors = append(errors, fmt.Errorf("Ingress %s/%s: Invalid value for nginx.org/limit-req-zone-size: got %s: %w", context.GetNamespace(), context.GetName(), requestRateZoneSize, err))
+		} else {
+			cfgParams.LimitReqZoneSize = size
+		}
+	}
+	if requestRateDelay, exists, err := GetMapKeyAsInt(annotations, "nginx.org/limit-req-delay", context); exists {
+		if err != nil {
+			errors = append(errors, err)
+		} else {
+			cfgParams.LimitReqDelay = requestRateDelay
+		}
+	}
+	if requestRateNoDelay, exists, err := GetMapKeyAsBool(annotations, "nginx.org/limit-req-no-delay", context); exists {
+		if err != nil {
+			errors = append(errors, err)
+		} else {
+			cfgParams.LimitReqNoDelay = requestRateNoDelay
+		}
+	}
+	if requestRateBurst, exists, err := GetMapKeyAsInt(annotations, "nginx.org/limit-req-burst", context); exists {
+		if err != nil {
+			errors = append(errors, err)
+		} else {
+			cfgParams.LimitReqBurst = requestRateBurst
+		}
+	}
+	if requestRateDryRun, exists, err := GetMapKeyAsBool(annotations, "nginx.org/limit-req-dry-run", context); exists {
+		if err != nil {
+			errors = append(errors, err)
+		} else {
+			cfgParams.LimitReqDryRun = requestRateDryRun
+		}
+	}
+	if requestRateLogLevel, exists := annotations["nginx.org/limit-req-log-level"]; exists {
+		if !slices.Contains([]string{"info", "notice", "warn", "error"}, requestRateLogLevel) {
+			errors = append(errors, fmt.Errorf("Ingress %s/%s: Invalid value for nginx.org/limit-req-log-level: got %s", context.GetNamespace(), context.GetName(), requestRateLogLevel))
+		} else {
+			cfgParams.LimitReqLogLevel = requestRateLogLevel
+		}
+	}
+	if requestRateRejectCode, exists, err := GetMapKeyAsInt(annotations, "nginx.org/limit-req-reject-code", context); exists {
+		if err != nil {
+			errors = append(errors, err)
+		} else {
+			cfgParams.LimitReqRejectCode = requestRateRejectCode
+		}
+	}
+	return errors
+}
+
 func getWebsocketServices(ingEx *IngressEx) map[string]bool {
 	if value, exists := ingEx.Ingress.Annotations["nginx.org/websocket-services"]; exists {
 		return ParseServiceList(value)

internal/configs/annotations_test.go

@@ -4,6 +4,9 @@ import (
 	"reflect"
 	"sort"
 	"testing"
+
+	networking "k8s.io/api/networking/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 )
 
 func TestParseRewrites(t *testing.T) {
@@ -159,3 +162,56 @@ func TestMergeMasterAnnotationsIntoMinion(t *testing.T) {
 		t.Errorf("mergeMasterAnnotationsIntoMinion returned %v, but expected %v", minionAnnotations, expectedMergedAnnotations)
 	}
 }
+
+func TestParseRateLimitAnnotations(t *testing.T) {
+	context := &networking.Ingress{
+		ObjectMeta: metav1.ObjectMeta{
+			Namespace: "default",
+			Name:      "context",
+		},
+	}
+
+	if errors := parseRateLimitAnnotations(map[string]string{
+		"nginx.org/limit-req-rate":        "200r/s",
+		"nginx.org/limit-req-key":         "${request_uri}",
+		"nginx.org/limit-req-burst":       "100",
+		"nginx.org/limit-req-delay":       "80",
+		"nginx.org/limit-req-no-delay":    "true",
+		"nginx.org/limit-req-reject-code": "429",
+		"nginx.org/limit-req-zone-size":   "11m",
+		"nginx.org/limit-req-dry-run":     "true",
+		"nginx.org/limit-req-log-level":   "info",
+	}, NewDefaultConfigParams(false), context); len(errors) > 0 {
+		t.Error("Errors when parsing valid limit-req annotations")
+	}
+
+	if errors := parseRateLimitAnnotations(map[string]string{
+		"nginx.org/limit-req-rate": "200",
+	}, NewDefaultConfigParams(false), context); len(errors) == 0 {
+		t.Error("No Errors when parsing invalid request rate")
+	}
+
+	if errors := parseRateLimitAnnotations(map[string]string{
+		"nginx.org/limit-req-rate": "200r/h",
+	}, NewDefaultConfigParams(false), context); len(errors) == 0 {
+		t.Error("No Errors when parsing invalid request rate")
+	}
+
+	if errors := parseRateLimitAnnotations(map[string]string{
+		"nginx.org/limit-req-rate": "0r/s",
+	}, NewDefaultConfigParams(false), context); len(errors) == 0 {
+		t.Error("No Errors when parsing invalid request rate")
+	}
+
+	if errors := parseRateLimitAnnotations(map[string]string{
+		"nginx.org/limit-req-zone-size": "10abc",
+	}, NewDefaultConfigParams(false), context); len(errors) == 0 {
+		t.Error("No Errors when parsing invalid zone size")
+	}
+
+	if errors := parseRateLimitAnnotations(map[string]string{
+		"nginx.org/limit-req-log-level": "foobar",
+	}, NewDefaultConfigParams(false), context); len(errors) == 0 {
+		t.Error("No Errors when parsing invalid log level")
+	}
+}

internal/configs/config_params.go

@@ -113,6 +113,16 @@ type ConfigParams struct {
 	SSLPorts []int
 
 	SpiffeServerCerts bool
+
+	LimitReqRate       string
+	LimitReqKey        string
+	LimitReqZoneSize   string
+	LimitReqDelay      int
+	LimitReqNoDelay    bool
+	LimitReqBurst      int
+	LimitReqDryRun     bool
+	LimitReqLogLevel   string
+	LimitReqRejectCode int
 }
 
 // StaticConfigParams holds immutable NGINX configuration parameters that affect the main NGINX config.
@@ -191,6 +201,10 @@ func NewDefaultConfigParams(isPlus bool) *ConfigParams {
 		MainKeepaliveRequests:         100,
 		VariablesHashBucketSize:       256,
 		VariablesHashMaxSize:          1024,
+		LimitReqKey:                   "${binary_remote_addr}",
+		LimitReqZoneSize:              "10m",
+		LimitReqLogLevel:              "error",
+		LimitReqRejectCode:            429,
 	}
 }
 

internal/configs/ingress.go

@@ -128,6 +128,7 @@ func generateNginxCfg(p NginxCfgParams) (version1.IngressNginxConfig, Warnings)
 	allWarnings := newWarnings()
 
 	var servers []version1.Server
+	var limitReqZones []version1.LimitReqZone
 
 	for _, rule := range p.ingEx.Ingress.Spec.Rules {
 		// skipping invalid hosts
@@ -265,6 +266,27 @@ func generateNginxCfg(p NginxCfgParams) (version1.IngressNginxConfig, Warnings)
 				allWarnings.Add(warnings)
 			}
 
+			if cfgParams.LimitReqRate != "" {
+				zoneName := p.ingEx.Ingress.Namespace + "/" + p.ingEx.Ingress.Name
+				loc.LimitReq = &version1.LimitReq{
+					Zone:       zoneName,
+					Burst:      cfgParams.LimitReqBurst,
+					Delay:      cfgParams.LimitReqDelay,
+					NoDelay:    cfgParams.LimitReqNoDelay,
+					DryRun:     cfgParams.LimitReqDryRun,
+					LogLevel:   cfgParams.LimitReqLogLevel,
+					RejectCode: cfgParams.LimitReqRejectCode,
+				}
+				if !limitReqZoneExists(limitReqZones, zoneName) {
+					limitReqZones = append(limitReqZones, version1.LimitReqZone{
+						Name: zoneName,
+						Key:  cfgParams.LimitReqKey,
+						Size: cfgParams.LimitReqZoneSize,
+						Rate: cfgParams.LimitReqRate,
+					})
+				}
+			}
+
 			locations = append(locations, loc)
 
 			if loc.Path == "/" {
@@ -317,6 +339,7 @@ func generateNginxCfg(p NginxCfgParams) (version1.IngressNginxConfig, Warnings)
 		SpiffeClientCerts:       p.staticParams.NginxServiceMesh && !cfgParams.SpiffeServerCerts,
 		DynamicSSLReloadEnabled: p.staticParams.DynamicSSLReload,
 		StaticSSLPath:           p.staticParams.StaticSSLPath,
+		LimitReqZones:           limitReqZones,
 	}, allWarnings
 }
 
@@ -609,6 +632,7 @@ func generateNginxCfgForMergeableIngresses(p NginxCfgParams) (version1.IngressNg
 	var locations []version1.Location
 	var upstreams []version1.Upstream
 	healthChecks := make(map[string]version1.HealthCheck)
+	var limitReqZones []version1.LimitReqZone
 	var keepalive string
 
 	// replace master with a deepcopy because we will modify it
@@ -704,6 +728,7 @@ func generateNginxCfgForMergeableIngresses(p NginxCfgParams) (version1.IngressNg
 		}
 
 		upstreams = append(upstreams, nginxCfg.Upstreams...)
+		limitReqZones = append(limitReqZones, nginxCfg.LimitReqZones...)
 	}
 
 	masterServer.HealthChecks = healthChecks
@@ -717,9 +742,19 @@ func generateNginxCfgForMergeableIngresses(p NginxCfgParams) (version1.IngressNg
 		SpiffeClientCerts:       p.staticParams.NginxServiceMesh && !p.baseCfgParams.SpiffeServerCerts,
 		DynamicSSLReloadEnabled: p.staticParams.DynamicSSLReload,
 		StaticSSLPath:           p.staticParams.StaticSSLPath,
+		LimitReqZones:           limitReqZones,
 	}, warnings
 }
 
+func limitReqZoneExists(zones []version1.LimitReqZone, zoneName string) bool {
+	for _, zone := range zones {
+		if zone.Name == zoneName {
+			return true
+		}
+	}
+	return false
+}
+
 func isSSLEnabled(isSSLService bool, cfgParams ConfigParams, staticCfgParams *StaticConfigParams) bool {
 	return isSSLService || staticCfgParams.NginxServiceMesh && !cfgParams.SpiffeServerCerts
 }