Merge pull request #106366 from cyclinder/evictions_number_stable

adding evictions_total metric and marking evictions_number deprecated

pkg/controller/nodelifecycle/metrics.go

@@ -29,6 +29,7 @@ const (
 	zoneSizeKey             = "zone_size"
 	zoneNoUnhealthyNodesKey = "unhealthy_nodes_in_zone"
 	evictionsNumberKey      = "evictions_number"
+	evictionsTotalKey       = "evictions_total"
 )
 
 var (
@@ -60,11 +61,21 @@ var (
 		[]string{"zone"},
 	)
 	evictionsNumber = metrics.NewCounterVec(
+		&metrics.CounterOpts{
+			Subsystem:         nodeControllerSubsystem,
+			Name:              evictionsNumberKey,
+			Help:              "Number of Node evictions that happened since current instance of NodeController started, This metric is replaced by node_collector_evictions_total.",
+			DeprecatedVersion: "1.24.0",
+			StabilityLevel:    metrics.ALPHA,
+		},
+		[]string{"zone"},
+	)
+	evictionsTotal = metrics.NewCounterVec(
 		&metrics.CounterOpts{
 			Subsystem:      nodeControllerSubsystem,
-			Name:           evictionsNumberKey,
+			Name:           evictionsTotalKey,
 			Help:           "Number of Node evictions that happened since current instance of NodeController started.",
-			StabilityLevel: metrics.ALPHA,
+			StabilityLevel: metrics.STABLE,
 		},
 		[]string{"zone"},
 	)
@@ -79,5 +90,6 @@ func Register() {
 		legacyregistry.MustRegister(zoneSize)
 		legacyregistry.MustRegister(unhealthyNodes)
 		legacyregistry.MustRegister(evictionsNumber)
+		legacyregistry.MustRegister(evictionsTotal)
 	})
 }

pkg/controller/nodelifecycle/node_lifecycle_controller.go

@@ -699,6 +699,7 @@ func (nc *Controller) doNoExecuteTaintingPass(ctx context.Context) {
 				//count the evictionsNumber
 				zone := nodetopology.GetZoneKey(node)
 				evictionsNumber.WithLabelValues(zone).Inc()
+				evictionsTotal.WithLabelValues(zone).Inc()
 			}
 
 			return result, 0
@@ -742,6 +743,7 @@ func (nc *Controller) doEvictionPass(ctx context.Context) {
 			if node != nil {
 				zone := nodetopology.GetZoneKey(node)
 				evictionsNumber.WithLabelValues(zone).Inc()
+				evictionsTotal.WithLabelValues(zone).Inc()
 			}
 
 			return true, 0
@@ -1396,6 +1398,7 @@ func (nc *Controller) addPodEvictorForNewZone(node *v1.Node) {
 		// Init the metric for the new zone.
 		klog.Infof("Initializing eviction metric for zone: %v", zone)
 		evictionsNumber.WithLabelValues(zone).Add(0)
+		evictionsTotal.WithLabelValues(zone).Add(0)
 	}
 }
 

test/instrumentation/testdata/stable-metrics-list.yaml

@@ -1,3 +1,11 @@
+- name: evictions_total
+  subsystem: node_collector
+  help: Number of Node evictions that happened since current instance of NodeController
+    started.
+  type: Counter
+  stabilityLevel: STABLE
+  labels:
+  - zone
 - name: framework_extension_point_duration_seconds
   subsystem: scheduler
   help: Latency for running all plugins of a specific extension point.