fix: replace os.Exit in goroutine with context cancellation for graceful shutdown

The goroutine in main.go that waits for webhook readiness and sets up
controllers was calling os.Exit(1) on failure, which kills the process
immediately without releasing the leader election lease. Other replicas
cannot acquire leadership until the lease TTL expires (default 15s).
Additionally, WaitReady() was an infinite loop with no context or
timeout, risking a zombie leader state if the webhook health check
never passes.
This commit:
- Wraps the signal handler context with context.WithCancel
- Replaces os.Exit(1) in the goroutine with cancel() + return,
  triggering the manager's graceful shutdown path
- Updates WaitReady to accept a context.Context so it can exit
  when the context is cancelled
Fixes #2401

Author: abhaygoudannavar

main.go

@@ -17,6 +17,7 @@ limitations under the License.
 package main
 
 import (
+	"context"
 	"flag"
 	"net/http"
 	_ "net/http/pprof"
@@ -164,7 +165,9 @@ func main() {
 		})
 	}
 
-	ctx := ctrl.SetupSignalHandler()
+	signalCtx := ctrl.SetupSignalHandler()
+	ctx, cancel := context.WithCancel(signalCtx)
+	defer cancel()
 	cfg := ctrl.GetConfigOrDie()
 	setRestConfig(cfg)
 	cfg.UserAgent = "kruise-manager"
@@ -256,15 +259,17 @@ func main() {
 
 	go func() {
 		setupLog.Info("wait webhook ready")
-		if err = webhook.WaitReady(); err != nil {
+		if err = webhook.WaitReady(ctx); err != nil {
 			setupLog.Error(err, "unable to wait webhook ready")
-			os.Exit(1)
+			cancel()
+			return
 		}
 
 		setupLog.Info("setup controllers")
 		if err = controller.SetupWithManager(mgr); err != nil {
 			setupLog.Error(err, "unable to setup controllers")
-			os.Exit(1)
+			cancel()
+			return
 		}
 	}()
 

pkg/webhook/server.go

@@ -133,10 +133,16 @@ func Checker(req *http.Request) error {
 	return health.Checker(req)
 }
 
-func WaitReady() error {
+func WaitReady(ctx context.Context) error {
 	startTS := time.Now()
 	var err error
 	for {
+		select {
+		case <-ctx.Done():
+			return fmt.Errorf("context cancelled while waiting for webhook ready: %v", ctx.Err())
+		default:
+		}
+
 		duration := time.Since(startTS)
 		if err = Checker(nil); err == nil {
 			return nil
@@ -147,5 +153,4 @@ func WaitReady() error {
 		}
 		time.Sleep(time.Second * 2)
 	}
-
 }