Merge pull request #685 from JaneLiuL/master

add  liveness probe

charts/descheduler/templates/deployment.yaml

@@ -47,6 +47,14 @@ spec:
           ports:
             - containerPort: 10258
               protocol: TCP
+          livenessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /healthz
+              port: 10258
+              scheme: HTTPS
+            initialDelaySeconds: 3
+            periodSeconds: 3
           resources:
             {{- toYaml .Values.resources | nindent 12 }}
           securityContext:

cmd/descheduler/app/server.go

@@ -20,6 +20,7 @@ package app
 import (
 	"context"
 	"io"
+	"k8s.io/apiserver/pkg/server/healthz"
 
 	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"sigs.k8s.io/descheduler/pkg/descheduler"
@@ -68,15 +69,17 @@ func NewDeschedulerCommand(out io.Writer) *cobra.Command {
 				klog.SetLogger(log)
 			}
 
+			ctx := context.TODO()
+			pathRecorderMux := mux.NewPathRecorderMux("descheduler")
 			if !s.DisableMetrics {
-				ctx := context.TODO()
-				pathRecorderMux := mux.NewPathRecorderMux("descheduler")
 				pathRecorderMux.Handle("/metrics", legacyregistry.HandlerWithReset())
+			}
 
-				if _, err := SecureServing.Serve(pathRecorderMux, 0, ctx.Done()); err != nil {
-					klog.Fatalf("failed to start secure server: %v", err)
-					return
-				}
+			healthz.InstallHandler(pathRecorderMux, healthz.NamedCheck("Descheduler", healthz.PingHealthz.Check))
+
+			if _, err := SecureServing.Serve(pathRecorderMux, 0, ctx.Done()); err != nil {
+				klog.Fatalf("failed to start secure server: %v", err)
+				return
 			}
 
 			err := Run(s)

kubernetes/cronjob/cronjob.yaml

@@ -31,6 +31,14 @@ spec:
               requests:
                 cpu: "500m"
                 memory: "256Mi"
+            livenessProbe:
+              failureThreshold: 3
+              httpGet:
+                path: /healthz
+                port: 10258
+                scheme: HTTPS
+              initialDelaySeconds: 3
+              periodSeconds: 10
             securityContext:
               allowPrivilegeEscalation: false
               capabilities:

kubernetes/deployment/deployment.yaml

@@ -33,6 +33,14 @@ spec:
           ports:
           - containerPort: 10258
             protocol: TCP
+          livenessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /healthz
+              port: 10258
+              scheme: HTTPS
+            initialDelaySeconds: 3
+            periodSeconds: 3
           resources:
             requests:
               cpu: 500m

kubernetes/job/job.yaml

@@ -29,6 +29,14 @@ spec:
             requests:
               cpu: "500m"
               memory: "256Mi"
+          livenessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /healthz
+              port: 10258
+              scheme: HTTPS
+            initialDelaySeconds: 3
+            periodSeconds: 3
           securityContext:
             allowPrivilegeEscalation: false
             capabilities: