Implement reporting status classes in metrics (#8548)

This commit introduces a backwards compatible command line option
--report-status-classes which will enable reporting response status classes
(2xx, 3xx..) instead of status codes in exported metrics.

cmd/nginx/flags.go

@@ -163,6 +163,9 @@ Requires the update-status parameter.`)
 			`Enables the collection of NGINX metrics`)
 		metricsPerHost = flags.Bool("metrics-per-host", true,
 			`Export metrics per-host`)
+		reportStatusClasses = flags.Bool("report-status-classes", false,
+			`Use status classes (2xx, 3xx, 4xx and 5xx) instead of status codes in metrics`)
+
 		timeBuckets         = flags.Float64Slice("time-buckets", prometheus.DefBuckets, "Set of buckets which will be used for prometheus histogram metrics such as RequestTime, ResponseTime")
 		lengthBuckets       = flags.Float64Slice("length-buckets", prometheus.LinearBuckets(10, 10, 10), "Set of buckets which will be used for prometheus histogram metrics such as RequestLength, ResponseLength")
 		sizeBuckets         = flags.Float64Slice("size-buckets", prometheus.ExponentialBuckets(10, 10, 7), "Set of buckets which will be used for prometheus histogram metrics such as BytesSent")
@@ -313,6 +316,7 @@ https://blog.maxmind.com/2019/12/18/significant-changes-to-accessing-and-using-g
 		EnableMetrics:               *enableMetrics,
 		MetricsPerHost:              *metricsPerHost,
 		MetricsBuckets:              histogramBuckets,
+		ReportStatusClasses:         *reportStatusClasses,
 		MonitorMaxBatchSize:         *monitorMaxBatchSize,
 		DisableServiceExternalName:  *disableServiceExternalName,
 		EnableSSLPassthrough:        *enableSSLPassthrough,

cmd/nginx/main.go

@@ -133,7 +133,7 @@ func main() {
 
 	mc := metric.NewDummyCollector()
 	if conf.EnableMetrics {
-		mc, err = metric.NewCollector(conf.MetricsPerHost, reg, conf.IngressClassConfiguration.Controller, *conf.MetricsBuckets)
+		mc, err = metric.NewCollector(conf.MetricsPerHost, conf.ReportStatusClasses, reg, conf.IngressClassConfiguration.Controller, *conf.MetricsBuckets)
 		if err != nil {
 			klog.Fatalf("Error creating prometheus collector:  %v", err)
 		}

docs/user-guide/cli-arguments.md

@@ -47,6 +47,7 @@ They are set in the container spec of the `ingress-nginx-controller` Deployment
 | `--publish-service`                | Service fronting the Ingress controller. Takes the form "namespace/name". When used together with update-status, the controller mirrors the address of this service's endpoints to the load-balancer status of all Ingress objects it satisfies. |
 | `--publish-status-address`         | Customized address (or addresses, separated by comma) to set as the load-balancer status of Ingress objects this controller satisfies. Requires the update-status parameter. |
 | `--report-node-internal-ip-address`| Set the load-balancer status of Ingress objects to internal Node addresses instead of external. Requires the update-status parameter. |
+| `--report-status-classes`          | If true, report status classes in metrics (2xx, 3xx, 4xx and 5xx) instead of full status codes. (default false) |
 | `--skip_headers`                   | If true, avoid header prefixes in the log messages |
 | `--skip_log_headers`               | If true, avoid headers when opening log files |
 | `--ssl-passthrough-proxy-port`     | Port to use internally for SSL Passthrough. (default 442) |

internal/ingress/controller/controller.go

@@ -100,6 +100,7 @@ type Configuration struct {
 	EnableMetrics       bool
 	MetricsPerHost      bool
 	MetricsBuckets      *collectors.HistogramBuckets
+	ReportStatusClasses bool
 
 	FakeCertificate *ingress.SSLCert
 

internal/ingress/metric/collectors/socket.go

@@ -86,6 +86,7 @@ type SocketCollector struct {
 	hosts sets.String
 
 	metricsPerHost      bool
+	reportStatusClasses bool
 
 	buckets HistogramBuckets
 }
@@ -110,7 +111,7 @@ var defObjectives = map[float64]float64{0.5: 0.05, 0.9: 0.01, 0.99: 0.001}
 
 // NewSocketCollector creates a new SocketCollector instance using
 // the ingress watch namespace and class used by the controller
-func NewSocketCollector(pod, namespace, class string, metricsPerHost bool, buckets HistogramBuckets) (*SocketCollector, error) {
+func NewSocketCollector(pod, namespace, class string, metricsPerHost, reportStatusClasses bool, buckets HistogramBuckets) (*SocketCollector, error) {
 	socket := "/tmp/nginx/prometheus-nginx.socket"
 	// unix sockets must be unlink()ed before being used
 	_ = syscall.Unlink(socket)
@@ -140,6 +141,7 @@ func NewSocketCollector(pod, namespace, class string, metricsPerHost bool, bucke
 		listener: listener,
 
 		metricsPerHost:      metricsPerHost,
+		reportStatusClasses: reportStatusClasses,
 
 		responseTime: prometheus.NewHistogramVec(
 			prometheus.HistogramOpts{
@@ -249,6 +251,10 @@ func (sc *SocketCollector) handleMessage(msg []byte) {
 			continue
 		}
 
+		if sc.reportStatusClasses && len(stats.Status) > 0 {
+			stats.Status = fmt.Sprintf("%cxx", stats.Status[0])
+		}
+
 		// Note these must match the order in requestTags at the top
 		requestLabels := prometheus.Labels{
 			"status":    stats.Status,

internal/ingress/metric/collectors/socket_test.go

@@ -83,6 +83,7 @@ func TestCollector(t *testing.T) {
 		name             string
 		data             []string
 		metrics          []string
+		useStatusClasses bool
 		wantBefore       string
 		removeIngresses  []string
 		wantAfter        string
@@ -371,13 +372,55 @@ func TestCollector(t *testing.T) {
 			wantAfter: `
 			`,
 		},
+		{
+			name: "valid metric object with status classes should update prometheus metrics",
+			data: []string{`[{
+				"host":"testshop.com",
+				"status":"200",
+				"bytesSent":150.0,
+				"method":"GET",
+				"path":"/admin",
+				"requestLength":300.0,
+				"requestTime":60.0,
+				"upstreamName":"test-upstream",
+				"upstreamIP":"1.1.1.1:8080",
+				"upstreamResponseTime":200,
+				"upstreamStatus":"220",
+				"namespace":"test-app-production",
+				"ingress":"web-yml",
+				"service":"test-app",
+				"canary":""
+			}]`},
+			metrics:          []string{"nginx_ingress_controller_response_duration_seconds"},
+			useStatusClasses: true,
+			wantBefore: `
+				# HELP nginx_ingress_controller_response_duration_seconds The time spent on receiving the response from the upstream server
+				# TYPE nginx_ingress_controller_response_duration_seconds histogram
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.005"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.01"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.025"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.05"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.1"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.25"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="0.5"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="1"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="2.5"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="5"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="10"} 0
+				nginx_ingress_controller_response_duration_seconds_bucket{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx",le="+Inf"} 1
+				nginx_ingress_controller_response_duration_seconds_sum{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx"} 200
+				nginx_ingress_controller_response_duration_seconds_count{canary="",controller_class="ingress",controller_namespace="default",controller_pod="pod",host="testshop.com",ingress="web-yml",method="GET",namespace="test-app-production",path="/admin",service="test-app",status="2xx"} 1
+			`,
+			wantAfter: `
+			`,
+		},
 	}
 
 	for _, c := range cases {
 		t.Run(c.name, func(t *testing.T) {
 			registry := prometheus.NewPedanticRegistry()
 
-			sc, err := NewSocketCollector("pod", "default", "ingress", true, buckets)
+			sc, err := NewSocketCollector("pod", "default", "ingress", true, c.useStatusClasses, buckets)
 			if err != nil {
 				t.Errorf("%v: unexpected error creating new SocketCollector: %v", c.name, err)
 			}

internal/ingress/metric/main.go

@@ -69,7 +69,7 @@ type collector struct {
 }
 
 // NewCollector creates a new metric collector the for ingress controller
-func NewCollector(metricsPerHost bool, registry *prometheus.Registry, ingressclass string, buckets collectors.HistogramBuckets) (Collector, error) {
+func NewCollector(metricsPerHost, reportStatusClasses bool, registry *prometheus.Registry, ingressclass string, buckets collectors.HistogramBuckets) (Collector, error) {
 	podNamespace := os.Getenv("POD_NAMESPACE")
 	if podNamespace == "" {
 		podNamespace = "default"
@@ -87,7 +87,7 @@ func NewCollector(metricsPerHost bool, registry *prometheus.Registry, ingresscla
 		return nil, err
 	}
 
-	s, err := collectors.NewSocketCollector(podName, podNamespace, ingressclass, metricsPerHost, buckets)
+	s, err := collectors.NewSocketCollector(podName, podNamespace, ingressclass, metricsPerHost, reportStatusClasses, buckets)
 	if err != nil {
 		return nil, err
 	}