add informer sync timeout

Signed-off-by: Eric Pickard <piceri@github.com>

Author: piceri

internal/controller/controller.go

@@ -38,6 +38,10 @@ const (
 	// deployment record API. Once an artifact is known to be missing,
 	// we suppress further API calls for this duration.
 	unknownArtifactTTL = 1 * time.Hour
+
+	// informerSyncTimeoutDuration is the maximum duration of time allowed
+	// for the informers to sync to prevent the controller from hanging indefinitely.
+	informerSyncTimeoutDuration = 60 * time.Second
 )
 
 type ttlCache interface {
@@ -92,6 +96,9 @@ type Controller struct {
 	// best effort cache to suppress API calls for artifacts that
 	// returned a 404 (no artifact found). Keyed by image digest.
 	unknownArtifacts ttlCache
+	// informerSyncTimeout is the maximum time allowed for all informers to sync
+	// and prevents sync from hanging indefinitely.
+	informerSyncTimeout time.Duration
 }
 
 // New creates a new deployment tracker controller.
@@ -160,6 +167,7 @@ func New(clientset kubernetes.Interface, metadataAggregator podMetadataAggregato
 		cfg:                 cfg,
 		observedDeployments: amcache.NewExpiring(),
 		unknownArtifacts:    amcache.NewExpiring(),
+		informerSyncTimeout: informerSyncTimeoutDuration,
 	}
 
 	// Add event handlers to the informer
@@ -320,16 +328,20 @@ func (c *Controller) Run(ctx context.Context, workers int) error {
 
 	// Wait for the caches to be synced
 	slog.Info("Waiting for informer caches to sync")
-	if !cache.WaitForCacheSync(ctx.Done(),
+	informerSyncCxt, cancel := context.WithTimeout(ctx, c.informerSyncTimeout)
+
+	if !cache.WaitForCacheSync(informerSyncCxt.Done(),
 		c.podInformer.HasSynced,
 		c.deploymentInformer.HasSynced,
 		c.daemonSetInformer.HasSynced,
 		c.statefulSetInformer.HasSynced,
 		c.jobInformer.HasSynced,
 		c.cronJobInformer.HasSynced,
 	) {
-		return errors.New("timed out waiting for caches to sync")
+		cancel()
+		return errors.New("timed out waiting for caches to sync - please ensure deployment tracker has the correct kubernetes permissions")
 	}
+	cancel()
 
 	slog.Info("Starting workers",
 		"count", workers,

internal/controller/controller_test.go

@@ -12,7 +12,11 @@ import (
 	"github.com/stretchr/testify/require"
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/runtime"
 	amcache "k8s.io/apimachinery/pkg/util/cache"
+	"k8s.io/client-go/kubernetes/fake"
+	k8stesting "k8s.io/client-go/testing"
+	"k8s.io/client-go/util/workqueue"
 )
 
 // mockPoster records all PostOne calls and returns a configurable error.
@@ -533,3 +537,46 @@ func TestIsTerminalPhase(t *testing.T) {
 		})
 	}
 }
+
+func TestRun_InformerSyncTimeout(t *testing.T) {
+	t.Parallel()
+	fakeClient := fake.NewSimpleClientset()
+	fakeClient.PrependReactor("list", "*", func(action k8stesting.Action) (bool, runtime.Object, error) {
+		// Block until the test context is cancelled.
+		<-make(chan struct{})
+		return true, nil, nil
+	})
+
+	factory := createInformerFactory(fakeClient, "", "")
+
+	ctrl := &Controller{
+		clientset:           fakeClient,
+		podInformer:         factory.Core().V1().Pods().Informer(),
+		deploymentInformer:  factory.Apps().V1().Deployments().Informer(),
+		daemonSetInformer:   factory.Apps().V1().DaemonSets().Informer(),
+		statefulSetInformer: factory.Apps().V1().StatefulSets().Informer(),
+		jobInformer:         factory.Batch().V1().Jobs().Informer(),
+		cronJobInformer:     factory.Batch().V1().CronJobs().Informer(),
+		workqueue: workqueue.NewTypedRateLimitingQueue(
+			workqueue.DefaultTypedControllerRateLimiter[PodEvent](),
+		),
+		apiClient:           &mockPoster{},
+		cfg:                 &Config{},
+		observedDeployments: amcache.NewExpiring(),
+		unknownArtifacts:    amcache.NewExpiring(),
+		informerSyncTimeout: 2 * time.Second,
+	}
+
+	errCh := make(chan error, 1)
+	go func() {
+		errCh <- ctrl.Run(context.Background(), 1)
+	}()
+
+	select {
+	case err := <-errCh:
+		require.Error(t, err)
+		assert.Contains(t, err.Error(), "timed out waiting for caches to sync")
+	case <-time.After(5 * time.Second):
+		t.Fatal("Run did not return within 5 seconds — informer sync timeout was 2 seconds")
+	}
+}