feat: add --kube-api-cache-sync-timeout flag for configurable cache sync timeout

docs/flags.md

@@ -193,6 +193,7 @@ tags:
 | `--kubeconfig=""`                                                  | Retrieve target cluster configuration from a Kubernetes configuration file (default: auto-detect)                                                                                                                                                                                                                                                                                                                                                                                      |
 | `--request-timeout=30s`                                            | [DEPRECATED: use --kube-api-request-timeout] Request timeout when calling Kubernetes APIs. 0s means no timeout                                                                                                                                                                                                                                                                                                                                                                         |
 | `--kube-api-request-timeout=30s`                                   | Request timeout when calling Kubernetes APIs. 0s means no timeout                                                                                                                                                                                                                                                                                                                                                                                                                      |
+| `--kube-api-cache-sync-timeout=1m0s`                               | Timeout for waiting for Kubernetes informer caches to sync during startup. Values <= 0 use the default (60s). Increase only after ruling out RBAC, network, or API server issues.                                                                                                                                                                                                                                                                                                      |
 | `--kube-api-qps=5`                                                 | Maximum QPS to the Kubernetes API server from this client.                                                                                                                                                                                                                                                                                                                                                                                                                             |
 | `--kube-api-burst=10`                                              | Maximum burst for throttle to the Kubernetes API server from this client.                                                                                                                                                                                                                                                                                                                                                                                                              |
 | `--provider=provider`                                              | The DNS provider where the DNS records will be created (required, options: akamai, alibabacloud, aws, aws-sd, azure, azure-dns, azure-private-dns, civo, cloudflare, coredns, dnsimple, exoscale, gandi, godaddy, google, inmemory, linode, ns1, oci, ovh, pdns, pihole, plural, rfc2136, scaleway, skydns, transip, webhook)                                                                                                                                                          |

pkg/apis/externaldns/types.go

@@ -47,6 +47,7 @@ type Config struct {
 	APIServerURL                                  string
 	KubeConfig                                    string
 	RequestTimeout                                time.Duration
+	CacheSyncTimeout                              time.Duration
 	KubeAPIRequestTimeout                         time.Duration
 	KubeAPIQPS                                    int
 	KubeAPIBurst                                  int
@@ -356,6 +357,7 @@ var defaultConfig = &Config{
 	RegexDomainFilter:            regexp.MustCompile(""),
 	Registry:                     RegistryTXT,
 	RequestTimeout:               time.Second * 30,
+	CacheSyncTimeout:             time.Second * 60,
 	KubeAPIRequestTimeout:        time.Second * 30,
 	KubeAPIQPS:                   int(rest.DefaultQPS),
 	KubeAPIBurst:                 rest.DefaultBurst,
@@ -744,6 +746,7 @@ func bindFlags(b flags.FlagBinder, cfg *Config) {
 	b.StringVar("kubeconfig", "Retrieve target cluster configuration from a Kubernetes configuration file (default: auto-detect)", defaultConfig.KubeConfig, &cfg.KubeConfig)
 	b.DurationVar("request-timeout", "[DEPRECATED: use --kube-api-request-timeout] Request timeout when calling Kubernetes APIs. 0s means no timeout", defaultConfig.RequestTimeout, &cfg.RequestTimeout)
 	b.DurationVar("kube-api-request-timeout", "Request timeout when calling Kubernetes APIs. 0s means no timeout", defaultConfig.KubeAPIRequestTimeout, &cfg.KubeAPIRequestTimeout)
+	b.DurationVar("kube-api-cache-sync-timeout", "Timeout for waiting for Kubernetes informer caches to sync during startup. Values <= 0 use the default (60s). Increase only after ruling out RBAC, network, or API server issues.", defaultConfig.CacheSyncTimeout, &cfg.CacheSyncTimeout)
 	b.IntVar("kube-api-qps", "Maximum QPS to the Kubernetes API server from this client.", defaultConfig.KubeAPIQPS, &cfg.KubeAPIQPS)
 	b.IntVar("kube-api-burst", "Maximum burst for throttle to the Kubernetes API server from this client.", defaultConfig.KubeAPIBurst, &cfg.KubeAPIBurst)
 }

pkg/apis/externaldns/types_test.go

@@ -37,6 +37,7 @@ var (
 		APIServerURL:                           "",
 		KubeConfig:                             "",
 		RequestTimeout:                         time.Second * 30,
+		CacheSyncTimeout:                       time.Second * 60,
 		KubeAPIRequestTimeout:                  time.Second * 30,
 		KubeAPIQPS:                             int(rest.DefaultQPS),
 		KubeAPIBurst:                           rest.DefaultBurst,
@@ -144,6 +145,7 @@ var (
 		APIServerURL:                           "http://127.0.0.1:8080",
 		KubeConfig:                             "/some/path",
 		RequestTimeout:                         time.Second * 77,
+		CacheSyncTimeout:                       time.Second * 60,
 		KubeAPIRequestTimeout:                  time.Second * 77,
 		KubeAPIQPS:                             int(rest.DefaultQPS),
 		KubeAPIBurst:                           rest.DefaultBurst,

source/ambassador_host.go

@@ -96,7 +96,7 @@ func NewAmbassadorHostSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/contour_httpproxy.go

@@ -76,7 +76,7 @@ func NewContourHTTPProxySource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/f5_transportserver.go

@@ -79,7 +79,7 @@ func NewF5TransportServerSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/f5_virtualserver.go

@@ -78,7 +78,7 @@ func NewF5VirtualServerSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/gateway.go

@@ -214,18 +214,18 @@ func newGatewayRouteSource(
 	if rtInformerFactory != gwInformerFactory {
 		rtInformerFactory.Start(ctx.Done())
 	}
-	if err := informers.WaitForCacheSync(ctx, gwInformerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, gwInformerFactory, config.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 	if lsInformer != nil && lsInformerFactory != gwInformerFactory {
-		if err := informers.WaitForCacheSync(ctx, lsInformerFactory); err != nil {
+		if err := informers.WaitForCacheSync(ctx, lsInformerFactory, config.CacheSyncTimeout); err != nil {
 			return nil, err
 		}
 	}
-	if err := informers.WaitForCacheSync(ctx, rtInformerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, rtInformerFactory, config.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
-	if err := informers.WaitForCacheSync(ctx, kubeInformerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, kubeInformerFactory, config.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/gloo_proxy.go

@@ -186,10 +186,10 @@ func NewGlooSource(
 
 	informerFactory.Start(ctx.Done())
 	dynamicInformerFactory.Start(ctx.Done())
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
-	if err := informers.WaitForDynamicCacheSync(ctx, dynamicInformerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, dynamicInformerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/informers/informers.go

@@ -26,11 +26,11 @@ import (
 )
 
 const (
-	// defaultTimeout is the maximum time in seconds to wait for informer caches
-	// to complete initial sync. This is intentionally longer than the per-request
-	// timeout: a cache sync may require multiple sequential API calls
+	// DefaultCacheSyncTimeout is the maximum time in seconds to wait for informer
+	// caches to complete initial sync. This is intentionally longer than the
+	// per-request timeout: a cache sync may require multiple sequential API calls
 	// (LIST + Watch handshake), so the total wait needs to exceed a single request duration.
-	defaultTimeout = 60
+	DefaultCacheSyncTimeout = 60
 )
 
 type informerFactory interface {
@@ -41,21 +41,21 @@ type dynamicInformerFactory interface {
 	WaitForCacheSync(stopCh <-chan struct{}) map[schema.GroupVersionResource]bool
 }
 
-func WaitForCacheSync(ctx context.Context, factory informerFactory) error {
-	return waitForCacheSync(ctx, factory.WaitForCacheSync)
+func WaitForCacheSync(ctx context.Context, factory informerFactory, timeout time.Duration) error {
+	return waitForCacheSync(ctx, factory.WaitForCacheSync, timeout)
 }
 
-func WaitForDynamicCacheSync(ctx context.Context, factory dynamicInformerFactory) error {
-	return waitForCacheSync(ctx, factory.WaitForCacheSync)
+func WaitForDynamicCacheSync(ctx context.Context, factory dynamicInformerFactory, timeout time.Duration) error {
+	return waitForCacheSync(ctx, factory.WaitForCacheSync, timeout)
 }
 
-// waitForCacheSync waits for informer caches to sync with a default timeout.
+// waitForCacheSync waits for informer caches to sync within the given timeout.
+// If timeout is <= 0, the default is used.
 // Returns an error if any cache fails to sync, wrapping the context error if a timeout occurred.
-func waitForCacheSync[K comparable](ctx context.Context, waitFunc func(<-chan struct{}) map[K]bool) error {
-	// The function receives a ctx but then creates a new timeout,
-	// effectively overriding whatever deadline the caller may have set.
-	// If the caller passed a context with a 30s timeout, this function ignores it and waits 60s anyway.
-	timeout := defaultTimeout * time.Second
+func waitForCacheSync[K comparable](ctx context.Context, waitFunc func(<-chan struct{}) map[K]bool, timeout time.Duration) error {
+	if timeout <= 0 {
+		timeout = DefaultCacheSyncTimeout * time.Second
+	}
 	ctx, cancel := context.WithTimeout(ctx, timeout)
 	defer cancel()
 	for typ, done := range waitFunc(ctx.Done()) {

source/informers/informers_test.go

@@ -70,7 +70,7 @@ func TestWaitForCacheSync(t *testing.T) {
 			ctx := t.Context()
 
 			factory := &mockInformerFactory{syncResults: tt.syncResults}
-			err := WaitForCacheSync(ctx, factory)
+			err := WaitForCacheSync(ctx, factory, 0)
 
 			if tt.expectError {
 				assert.Error(t, err)
@@ -112,7 +112,7 @@ func TestWaitForDynamicCacheSync(t *testing.T) {
 			ctx := t.Context()
 
 			factory := &mockDynamicInformerFactory{syncResults: tt.syncResults}
-			err := WaitForDynamicCacheSync(ctx, factory)
+			err := WaitForDynamicCacheSync(ctx, factory, 0)
 
 			if tt.expectError {
 				assert.Error(t, err)

source/informers/transformers_test.go

@@ -324,7 +324,7 @@ func TestTransformerWithOptions_WithFakeClient(t *testing.T) {
 	require.NoError(t, err)
 
 	factory.Start(ctx.Done())
-	err = WaitForCacheSync(ctx, factory)
+	err = WaitForCacheSync(ctx, factory, 0)
 	require.NoError(t, err)
 
 	got, err := serviceInformer.Lister().Services(svc.Namespace).Get(svc.Name)

source/ingress.go

@@ -110,7 +110,7 @@ func NewIngressSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/istio_gateway.go

@@ -111,10 +111,10 @@ func NewIstioGatewaySource(
 	istioInformerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
-	if err := informers.WaitForCacheSync(ctx, istioInformerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, istioInformerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/istio_virtualservice.go

@@ -119,10 +119,10 @@ func NewIstioVirtualServiceSource(
 	istioInformerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
-	if err := informers.WaitForCacheSync(ctx, istioInformerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, istioInformerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/kong_tcpingress.go

@@ -84,7 +84,7 @@ func NewKongTCPIngressSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/node.go

@@ -80,7 +80,7 @@ func NewNodeSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/openshift_route.go

@@ -79,7 +79,7 @@ func NewOcpRouteSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/pod.go

@@ -96,7 +96,7 @@ func NewPodSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/service.go

@@ -159,7 +159,7 @@ func NewServiceSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForCacheSync(ctx, informerFactory, config.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/store.go

@@ -86,6 +86,7 @@ type Config struct {
 	GlooNamespaces                 []string
 	SkipperRouteGroupVersion       string
 	KubeAPIRequestTimeout          time.Duration
+	CacheSyncTimeout               time.Duration
 	KubeAPIQPS                     int
 	KubeAPIBurst                   int
 	DefaultTargets                 []string
@@ -161,6 +162,7 @@ func NewSourceConfig(cfg *externaldns.Config, opts ...OverrideConfigOption) (*Co
 		GlooNamespaces:                 cfg.GlooNamespaces,
 		SkipperRouteGroupVersion:       cfg.SkipperRouteGroupVersion,
 		KubeAPIRequestTimeout:          cfg.KubeAPIRequestTimeout,
+		CacheSyncTimeout:               cfg.CacheSyncTimeout,
 		KubeAPIQPS:                     cfg.KubeAPIQPS,
 		KubeAPIBurst:                   cfg.KubeAPIBurst,
 		DefaultTargets:                 cfg.DefaultTargets,

source/traefik_proxy.go

@@ -137,7 +137,7 @@ func NewTraefikSource(
 	informerFactory.Start(ctx.Done())
 
 	// wait for the local cache to be populated.
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}
 

source/unstructured.go

@@ -99,7 +99,7 @@ func NewUnstructuredFQDNSource(
 	}
 
 	informerFactory.Start(ctx.Done())
-	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory); err != nil {
+	if err := informers.WaitForDynamicCacheSync(ctx, informerFactory, cfg.CacheSyncTimeout); err != nil {
 		return nil, err
 	}