mirror of
				https://github.com/traefik/traefik.git
				synced 2025-10-31 00:11:38 +01:00 
			
		
		
		
	
		
			
				
	
	
		
			920 lines
		
	
	
		
			32 KiB
		
	
	
	
		
			Go
		
	
	
	
	
	
			
		
		
	
	
			920 lines
		
	
	
		
			32 KiB
		
	
	
	
		
			Go
		
	
	
	
	
	
| package gateway
 | |
| 
 | |
| import (
 | |
| 	"context"
 | |
| 	"errors"
 | |
| 	"fmt"
 | |
| 	"os"
 | |
| 	"reflect"
 | |
| 	"slices"
 | |
| 	"time"
 | |
| 
 | |
| 	"github.com/rs/zerolog/log"
 | |
| 	"github.com/traefik/traefik/v3/pkg/types"
 | |
| 	corev1 "k8s.io/api/core/v1"
 | |
| 	discoveryv1 "k8s.io/api/discovery/v1"
 | |
| 	kerror "k8s.io/apimachinery/pkg/api/errors"
 | |
| 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 | |
| 	"k8s.io/apimachinery/pkg/labels"
 | |
| 	"k8s.io/apimachinery/pkg/selection"
 | |
| 	ktypes "k8s.io/apimachinery/pkg/types"
 | |
| 	kinformers "k8s.io/client-go/informers"
 | |
| 	kclientset "k8s.io/client-go/kubernetes"
 | |
| 	"k8s.io/client-go/rest"
 | |
| 	"k8s.io/client-go/tools/clientcmd"
 | |
| 	"k8s.io/client-go/util/retry"
 | |
| 	gatev1 "sigs.k8s.io/gateway-api/apis/v1"
 | |
| 	gatev1alpha2 "sigs.k8s.io/gateway-api/apis/v1alpha2"
 | |
| 	gatev1alpha3 "sigs.k8s.io/gateway-api/apis/v1alpha3"
 | |
| 	gatev1beta1 "sigs.k8s.io/gateway-api/apis/v1beta1"
 | |
| 	gateclientset "sigs.k8s.io/gateway-api/pkg/client/clientset/versioned"
 | |
| 	gateinformers "sigs.k8s.io/gateway-api/pkg/client/informers/externalversions"
 | |
| )
 | |
| 
 | |
| const resyncPeriod = 10 * time.Minute
 | |
| 
 | |
| type resourceEventHandler struct {
 | |
| 	ev chan<- interface{}
 | |
| }
 | |
| 
 | |
| func (reh *resourceEventHandler) OnAdd(obj interface{}, _ bool) {
 | |
| 	eventHandlerFunc(reh.ev, obj)
 | |
| }
 | |
| 
 | |
| func (reh *resourceEventHandler) OnUpdate(_, newObj interface{}) {
 | |
| 	eventHandlerFunc(reh.ev, newObj)
 | |
| }
 | |
| 
 | |
| func (reh *resourceEventHandler) OnDelete(obj interface{}) {
 | |
| 	eventHandlerFunc(reh.ev, obj)
 | |
| }
 | |
| 
 | |
| type clientWrapper struct {
 | |
| 	csGateway gateclientset.Interface
 | |
| 	csKube    kclientset.Interface
 | |
| 
 | |
| 	factoryNamespace    kinformers.SharedInformerFactory
 | |
| 	factoryGatewayClass gateinformers.SharedInformerFactory
 | |
| 	factoriesGateway    map[string]gateinformers.SharedInformerFactory
 | |
| 	factoriesKube       map[string]kinformers.SharedInformerFactory
 | |
| 	factoriesSecret     map[string]kinformers.SharedInformerFactory
 | |
| 
 | |
| 	isNamespaceAll    bool
 | |
| 	watchedNamespaces []string
 | |
| 
 | |
| 	labelSelector       string
 | |
| 	experimentalChannel bool
 | |
| }
 | |
| 
 | |
| func createClientFromConfig(c *rest.Config) (*clientWrapper, error) {
 | |
| 	csGateway, err := gateclientset.NewForConfig(c)
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 
 | |
| 	csKube, err := kclientset.NewForConfig(c)
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 
 | |
| 	return newClientImpl(csKube, csGateway), nil
 | |
| }
 | |
| 
 | |
| func newClientImpl(csKube kclientset.Interface, csGateway gateclientset.Interface) *clientWrapper {
 | |
| 	return &clientWrapper{
 | |
| 		csGateway:        csGateway,
 | |
| 		csKube:           csKube,
 | |
| 		factoriesGateway: make(map[string]gateinformers.SharedInformerFactory),
 | |
| 		factoriesKube:    make(map[string]kinformers.SharedInformerFactory),
 | |
| 		factoriesSecret:  make(map[string]kinformers.SharedInformerFactory),
 | |
| 	}
 | |
| }
 | |
| 
 | |
| // newInClusterClient returns a new Provider client that is expected to run
 | |
| // inside the cluster.
 | |
| func newInClusterClient(endpoint string) (*clientWrapper, error) {
 | |
| 	config, err := rest.InClusterConfig()
 | |
| 	if err != nil {
 | |
| 		return nil, fmt.Errorf("failed to create in-cluster configuration: %w", err)
 | |
| 	}
 | |
| 
 | |
| 	if endpoint != "" {
 | |
| 		config.Host = endpoint
 | |
| 	}
 | |
| 
 | |
| 	return createClientFromConfig(config)
 | |
| }
 | |
| 
 | |
| func newExternalClusterClientFromFile(file string) (*clientWrapper, error) {
 | |
| 	configFromFlags, err := clientcmd.BuildConfigFromFlags("", file)
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 	return createClientFromConfig(configFromFlags)
 | |
| }
 | |
| 
 | |
| // newExternalClusterClient returns a new Provider client that may run outside of the cluster.
 | |
| // The endpoint parameter must not be empty.
 | |
| func newExternalClusterClient(endpoint, caFilePath string, token types.FileOrContent) (*clientWrapper, error) {
 | |
| 	if endpoint == "" {
 | |
| 		return nil, errors.New("endpoint missing for external cluster client")
 | |
| 	}
 | |
| 
 | |
| 	tokenData, err := token.Read()
 | |
| 	if err != nil {
 | |
| 		return nil, fmt.Errorf("read token: %w", err)
 | |
| 	}
 | |
| 
 | |
| 	config := &rest.Config{
 | |
| 		Host:        endpoint,
 | |
| 		BearerToken: string(tokenData),
 | |
| 	}
 | |
| 
 | |
| 	if caFilePath != "" {
 | |
| 		caData, err := os.ReadFile(caFilePath)
 | |
| 		if err != nil {
 | |
| 			return nil, fmt.Errorf("failed to read CA file %s: %w", caFilePath, err)
 | |
| 		}
 | |
| 
 | |
| 		config.TLSClientConfig = rest.TLSClientConfig{CAData: caData}
 | |
| 	}
 | |
| 
 | |
| 	return createClientFromConfig(config)
 | |
| }
 | |
| 
 | |
| // WatchAll starts namespace-specific controllers for all relevant kinds.
 | |
| func (c *clientWrapper) WatchAll(namespaces []string, stopCh <-chan struct{}) (<-chan interface{}, error) {
 | |
| 	eventCh := make(chan interface{}, 1)
 | |
| 	eventHandler := &resourceEventHandler{ev: eventCh}
 | |
| 
 | |
| 	if len(namespaces) == 0 {
 | |
| 		namespaces = []string{metav1.NamespaceAll}
 | |
| 		c.isNamespaceAll = true
 | |
| 	}
 | |
| 
 | |
| 	c.watchedNamespaces = namespaces
 | |
| 
 | |
| 	notOwnedByHelm := func(opts *metav1.ListOptions) {
 | |
| 		opts.LabelSelector = "owner!=helm"
 | |
| 	}
 | |
| 
 | |
| 	labelSelectorOptions := func(options *metav1.ListOptions) {
 | |
| 		options.LabelSelector = c.labelSelector
 | |
| 	}
 | |
| 
 | |
| 	c.factoryNamespace = kinformers.NewSharedInformerFactory(c.csKube, resyncPeriod)
 | |
| 	_, err := c.factoryNamespace.Core().V1().Namespaces().Informer().AddEventHandler(eventHandler)
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 
 | |
| 	c.factoryGatewayClass = gateinformers.NewSharedInformerFactoryWithOptions(c.csGateway, resyncPeriod, gateinformers.WithTweakListOptions(labelSelectorOptions))
 | |
| 	_, err = c.factoryGatewayClass.Gateway().V1().GatewayClasses().Informer().AddEventHandler(eventHandler)
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 
 | |
| 	for _, ns := range namespaces {
 | |
| 		factoryKube := kinformers.NewSharedInformerFactoryWithOptions(c.csKube, resyncPeriod, kinformers.WithNamespace(ns))
 | |
| 		_, err = factoryKube.Core().V1().Services().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 		_, err = factoryKube.Discovery().V1().EndpointSlices().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 
 | |
| 		factoryGateway := gateinformers.NewSharedInformerFactoryWithOptions(c.csGateway, resyncPeriod, gateinformers.WithNamespace(ns))
 | |
| 		_, err = factoryGateway.Gateway().V1().Gateways().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 		_, err = factoryGateway.Gateway().V1().HTTPRoutes().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 		_, err = factoryGateway.Gateway().V1().GRPCRoutes().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 		_, err = factoryGateway.Gateway().V1beta1().ReferenceGrants().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 
 | |
| 		if c.experimentalChannel {
 | |
| 			_, err = factoryGateway.Gateway().V1alpha2().TCPRoutes().Informer().AddEventHandler(eventHandler)
 | |
| 			if err != nil {
 | |
| 				return nil, err
 | |
| 			}
 | |
| 			_, err = factoryGateway.Gateway().V1alpha2().TLSRoutes().Informer().AddEventHandler(eventHandler)
 | |
| 			if err != nil {
 | |
| 				return nil, err
 | |
| 			}
 | |
| 			_, err = factoryGateway.Gateway().V1alpha3().BackendTLSPolicies().Informer().AddEventHandler(eventHandler)
 | |
| 			if err != nil {
 | |
| 				return nil, err
 | |
| 			}
 | |
| 			_, err = factoryKube.Core().V1().ConfigMaps().Informer().AddEventHandler(eventHandler)
 | |
| 			if err != nil {
 | |
| 				return nil, err
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		factorySecret := kinformers.NewSharedInformerFactoryWithOptions(c.csKube, resyncPeriod, kinformers.WithNamespace(ns), kinformers.WithTweakListOptions(notOwnedByHelm))
 | |
| 		_, err = factorySecret.Core().V1().Secrets().Informer().AddEventHandler(eventHandler)
 | |
| 		if err != nil {
 | |
| 			return nil, err
 | |
| 		}
 | |
| 
 | |
| 		c.factoriesGateway[ns] = factoryGateway
 | |
| 		c.factoriesKube[ns] = factoryKube
 | |
| 		c.factoriesSecret[ns] = factorySecret
 | |
| 	}
 | |
| 
 | |
| 	c.factoryNamespace.Start(stopCh)
 | |
| 	c.factoryGatewayClass.Start(stopCh)
 | |
| 
 | |
| 	for _, ns := range namespaces {
 | |
| 		c.factoriesGateway[ns].Start(stopCh)
 | |
| 		c.factoriesKube[ns].Start(stopCh)
 | |
| 		c.factoriesSecret[ns].Start(stopCh)
 | |
| 	}
 | |
| 
 | |
| 	for t, ok := range c.factoryNamespace.WaitForCacheSync(stopCh) {
 | |
| 		if !ok {
 | |
| 			return nil, fmt.Errorf("timed out waiting for controller caches to sync %s", t.String())
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	for t, ok := range c.factoryGatewayClass.WaitForCacheSync(stopCh) {
 | |
| 		if !ok {
 | |
| 			return nil, fmt.Errorf("timed out waiting for controller caches to sync %s", t.String())
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	for _, ns := range namespaces {
 | |
| 		for t, ok := range c.factoriesGateway[ns].WaitForCacheSync(stopCh) {
 | |
| 			if !ok {
 | |
| 				return nil, fmt.Errorf("timed out waiting for controller caches to sync %s in namespace %q", t.String(), ns)
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		for t, ok := range c.factoriesKube[ns].WaitForCacheSync(stopCh) {
 | |
| 			if !ok {
 | |
| 				return nil, fmt.Errorf("timed out waiting for controller caches to sync %s in namespace %q", t.String(), ns)
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		for t, ok := range c.factoriesSecret[ns].WaitForCacheSync(stopCh) {
 | |
| 			if !ok {
 | |
| 				return nil, fmt.Errorf("timed out waiting for controller caches to sync %s in namespace %q", t.String(), ns)
 | |
| 			}
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	return eventCh, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListNamespaces(selector labels.Selector) ([]string, error) {
 | |
| 	ns, err := c.factoryNamespace.Core().V1().Namespaces().Lister().List(selector)
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 
 | |
| 	var namespaces []string
 | |
| 	for _, namespace := range ns {
 | |
| 		if !c.isWatchedNamespace(namespace.Name) {
 | |
| 			log.Warn().Msgf("Namespace %q is not within %q watched namespaces", selector, namespace)
 | |
| 			continue
 | |
| 		}
 | |
| 		namespaces = append(namespaces, namespace.Name)
 | |
| 	}
 | |
| 	return namespaces, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListHTTPRoutes() ([]*gatev1.HTTPRoute, error) {
 | |
| 	var httpRoutes []*gatev1.HTTPRoute
 | |
| 	for _, namespace := range c.watchedNamespaces {
 | |
| 		routes, err := c.factoriesGateway[c.lookupNamespace(namespace)].Gateway().V1().HTTPRoutes().Lister().HTTPRoutes(namespace).List(labels.Everything())
 | |
| 		if err != nil {
 | |
| 			return nil, fmt.Errorf("listing HTTP routes in namespace %s", namespace)
 | |
| 		}
 | |
| 
 | |
| 		httpRoutes = append(httpRoutes, routes...)
 | |
| 	}
 | |
| 
 | |
| 	return httpRoutes, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListGRPCRoutes() ([]*gatev1.GRPCRoute, error) {
 | |
| 	var grpcRoutes []*gatev1.GRPCRoute
 | |
| 	for _, namespace := range c.watchedNamespaces {
 | |
| 		routes, err := c.factoriesGateway[c.lookupNamespace(namespace)].Gateway().V1().GRPCRoutes().Lister().GRPCRoutes(namespace).List(labels.Everything())
 | |
| 		if err != nil {
 | |
| 			return nil, fmt.Errorf("listing GRPC routes in namespace %s", namespace)
 | |
| 		}
 | |
| 
 | |
| 		grpcRoutes = append(grpcRoutes, routes...)
 | |
| 	}
 | |
| 
 | |
| 	return grpcRoutes, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListTCPRoutes() ([]*gatev1alpha2.TCPRoute, error) {
 | |
| 	var tcpRoutes []*gatev1alpha2.TCPRoute
 | |
| 	for _, namespace := range c.watchedNamespaces {
 | |
| 		routes, err := c.factoriesGateway[c.lookupNamespace(namespace)].Gateway().V1alpha2().TCPRoutes().Lister().TCPRoutes(namespace).List(labels.Everything())
 | |
| 		if err != nil {
 | |
| 			return nil, fmt.Errorf("listing TCP routes in namespace %s", namespace)
 | |
| 		}
 | |
| 
 | |
| 		tcpRoutes = append(tcpRoutes, routes...)
 | |
| 	}
 | |
| 
 | |
| 	return tcpRoutes, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListTLSRoutes() ([]*gatev1alpha2.TLSRoute, error) {
 | |
| 	var tlsRoutes []*gatev1alpha2.TLSRoute
 | |
| 	for _, namespace := range c.watchedNamespaces {
 | |
| 		routes, err := c.factoriesGateway[c.lookupNamespace(namespace)].Gateway().V1alpha2().TLSRoutes().Lister().TLSRoutes(namespace).List(labels.Everything())
 | |
| 		if err != nil {
 | |
| 			return nil, fmt.Errorf("listing TLS routes in namespace %s", namespace)
 | |
| 		}
 | |
| 
 | |
| 		tlsRoutes = append(tlsRoutes, routes...)
 | |
| 	}
 | |
| 
 | |
| 	return tlsRoutes, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListReferenceGrants(namespace string) ([]*gatev1beta1.ReferenceGrant, error) {
 | |
| 	if !c.isWatchedNamespace(namespace) {
 | |
| 		return nil, fmt.Errorf("failed to get ReferenceGrants: namespace %s is not within watched namespaces", namespace)
 | |
| 	}
 | |
| 
 | |
| 	referenceGrants, err := c.factoriesGateway[c.lookupNamespace(namespace)].Gateway().V1beta1().ReferenceGrants().Lister().ReferenceGrants(namespace).List(labels.Everything())
 | |
| 	if err != nil {
 | |
| 		return nil, err
 | |
| 	}
 | |
| 
 | |
| 	return referenceGrants, nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListGateways() []*gatev1.Gateway {
 | |
| 	var result []*gatev1.Gateway
 | |
| 
 | |
| 	for ns, factory := range c.factoriesGateway {
 | |
| 		gateways, err := factory.Gateway().V1().Gateways().Lister().Gateways(ns).List(labels.Everything())
 | |
| 		if err != nil {
 | |
| 			log.Error().Err(err).Msgf("Failed to list Gateways in namespace %s", ns)
 | |
| 			continue
 | |
| 		}
 | |
| 		result = append(result, gateways...)
 | |
| 	}
 | |
| 
 | |
| 	return result
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) ListGatewayClasses() ([]*gatev1.GatewayClass, error) {
 | |
| 	return c.factoryGatewayClass.Gateway().V1().GatewayClasses().Lister().List(labels.Everything())
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateGatewayClassStatus(ctx context.Context, name string, status gatev1.GatewayClassStatus) error {
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentGatewayClass, err := c.factoryGatewayClass.Gateway().V1().GatewayClasses().Lister().Get(name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		if conditionsEqual(currentGatewayClass.Status.Conditions, status.Conditions) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentGatewayClass = currentGatewayClass.DeepCopy()
 | |
| 		currentGatewayClass.Status = status
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1().GatewayClasses().UpdateStatus(ctx, currentGatewayClass, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update GatewayClass %s status: %w", name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateGatewayStatus(ctx context.Context, gateway ktypes.NamespacedName, status gatev1.GatewayStatus) error {
 | |
| 	if !c.isWatchedNamespace(gateway.Namespace) {
 | |
| 		return fmt.Errorf("cannot update Gateway status %s/%s: namespace is not within watched namespaces", gateway.Namespace, gateway.Name)
 | |
| 	}
 | |
| 
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentGateway, err := c.factoriesGateway[c.lookupNamespace(gateway.Namespace)].Gateway().V1().Gateways().Lister().Gateways(gateway.Namespace).Get(gateway.Name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		if gatewayStatusEqual(currentGateway.Status, status) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentGateway = currentGateway.DeepCopy()
 | |
| 		currentGateway.Status = status
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1().Gateways(gateway.Namespace).UpdateStatus(ctx, currentGateway, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update Gateway %s/%s status: %w", gateway.Namespace, gateway.Name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateHTTPRouteStatus(ctx context.Context, route ktypes.NamespacedName, status gatev1.HTTPRouteStatus) error {
 | |
| 	if !c.isWatchedNamespace(route.Namespace) {
 | |
| 		return fmt.Errorf("updating HTTPRoute status %s/%s: namespace is not within watched namespaces", route.Namespace, route.Name)
 | |
| 	}
 | |
| 
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentRoute, err := c.factoriesGateway[c.lookupNamespace(route.Namespace)].Gateway().V1().HTTPRoutes().Lister().HTTPRoutes(route.Namespace).Get(route.Name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		parentStatuses := make([]gatev1.RouteParentStatus, len(status.Parents))
 | |
| 		copy(parentStatuses, status.Parents)
 | |
| 
 | |
| 		// keep statuses added by other gateway controllers.
 | |
| 		// TODO: we should also keep statuses for gateways managed by other Traefik instances.
 | |
| 		for _, parentStatus := range currentRoute.Status.Parents {
 | |
| 			if parentStatus.ControllerName != controllerName {
 | |
| 				parentStatuses = append(parentStatuses, parentStatus)
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		// do not update status when nothing has changed.
 | |
| 		if routeParentStatusesEqual(currentRoute.Status.Parents, parentStatuses) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentRoute = currentRoute.DeepCopy()
 | |
| 		currentRoute.Status = gatev1.HTTPRouteStatus{
 | |
| 			RouteStatus: gatev1.RouteStatus{
 | |
| 				Parents: parentStatuses,
 | |
| 			},
 | |
| 		}
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1().HTTPRoutes(route.Namespace).UpdateStatus(ctx, currentRoute, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update HTTPRoute %s/%s status: %w", route.Namespace, route.Name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateGRPCRouteStatus(ctx context.Context, route ktypes.NamespacedName, status gatev1.GRPCRouteStatus) error {
 | |
| 	if !c.isWatchedNamespace(route.Namespace) {
 | |
| 		return fmt.Errorf("updating GRPCRoute status %s/%s: namespace is not within watched namespaces", route.Namespace, route.Name)
 | |
| 	}
 | |
| 
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentRoute, err := c.factoriesGateway[c.lookupNamespace(route.Namespace)].Gateway().V1().GRPCRoutes().Lister().GRPCRoutes(route.Namespace).Get(route.Name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		parentStatuses := make([]gatev1.RouteParentStatus, len(status.Parents))
 | |
| 		copy(parentStatuses, status.Parents)
 | |
| 
 | |
| 		// keep statuses added by other gateway controllers.
 | |
| 		// TODO: we should also keep statuses for gateways managed by other Traefik instances.
 | |
| 		for _, parentStatus := range currentRoute.Status.Parents {
 | |
| 			if parentStatus.ControllerName != controllerName {
 | |
| 				parentStatuses = append(parentStatuses, parentStatus)
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		// do not update status when nothing has changed.
 | |
| 		if routeParentStatusesEqual(currentRoute.Status.Parents, parentStatuses) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentRoute = currentRoute.DeepCopy()
 | |
| 		currentRoute.Status = gatev1.GRPCRouteStatus{
 | |
| 			RouteStatus: gatev1.RouteStatus{
 | |
| 				Parents: parentStatuses,
 | |
| 			},
 | |
| 		}
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1().GRPCRoutes(route.Namespace).UpdateStatus(ctx, currentRoute, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update GRPCRoute %q status: %w", route.Name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateTCPRouteStatus(ctx context.Context, route ktypes.NamespacedName, status gatev1alpha2.TCPRouteStatus) error {
 | |
| 	if !c.isWatchedNamespace(route.Namespace) {
 | |
| 		return fmt.Errorf("updating TCPRoute status %s/%s: namespace is not within watched namespaces", route.Namespace, route.Name)
 | |
| 	}
 | |
| 
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentRoute, err := c.factoriesGateway[c.lookupNamespace(route.Namespace)].Gateway().V1alpha2().TCPRoutes().Lister().TCPRoutes(route.Namespace).Get(route.Name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		parentStatuses := make([]gatev1.RouteParentStatus, len(status.Parents))
 | |
| 		copy(parentStatuses, status.Parents)
 | |
| 
 | |
| 		// keep statuses added by other gateway controllers.
 | |
| 		// TODO: we should also keep statuses for gateways managed by other Traefik instances.
 | |
| 		for _, parentStatus := range currentRoute.Status.Parents {
 | |
| 			if parentStatus.ControllerName != controllerName {
 | |
| 				parentStatuses = append(parentStatuses, parentStatus)
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		// do not update status when nothing has changed.
 | |
| 		if routeParentStatusesEqual(currentRoute.Status.Parents, parentStatuses) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentRoute = currentRoute.DeepCopy()
 | |
| 		currentRoute.Status = gatev1alpha2.TCPRouteStatus{
 | |
| 			RouteStatus: gatev1.RouteStatus{
 | |
| 				Parents: parentStatuses,
 | |
| 			},
 | |
| 		}
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1alpha2().TCPRoutes(route.Namespace).UpdateStatus(ctx, currentRoute, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update TCPRoute %s/%s status: %w", route.Namespace, route.Name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateTLSRouteStatus(ctx context.Context, route ktypes.NamespacedName, status gatev1alpha2.TLSRouteStatus) error {
 | |
| 	if !c.isWatchedNamespace(route.Namespace) {
 | |
| 		return fmt.Errorf("updating TLSRoute status %s/%s: namespace is not within watched namespaces", route.Namespace, route.Name)
 | |
| 	}
 | |
| 
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentRoute, err := c.factoriesGateway[c.lookupNamespace(route.Namespace)].Gateway().V1alpha2().TLSRoutes().Lister().TLSRoutes(route.Namespace).Get(route.Name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		parentStatuses := make([]gatev1.RouteParentStatus, len(status.Parents))
 | |
| 		copy(parentStatuses, status.Parents)
 | |
| 
 | |
| 		// keep statuses added by other gateway controllers.
 | |
| 		// TODO: we should also keep statuses for gateways managed by other Traefik instances.
 | |
| 		for _, parentStatus := range currentRoute.Status.Parents {
 | |
| 			if parentStatus.ControllerName != controllerName {
 | |
| 				parentStatuses = append(parentStatuses, parentStatus)
 | |
| 			}
 | |
| 		}
 | |
| 
 | |
| 		// do not update status when nothing has changed.
 | |
| 		if routeParentStatusesEqual(currentRoute.Status.Parents, parentStatuses) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentRoute = currentRoute.DeepCopy()
 | |
| 		currentRoute.Status = gatev1alpha2.TLSRouteStatus{
 | |
| 			RouteStatus: gatev1.RouteStatus{
 | |
| 				Parents: parentStatuses,
 | |
| 			},
 | |
| 		}
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1alpha2().TLSRoutes(route.Namespace).UpdateStatus(ctx, currentRoute, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update TLSRoute %s/%s status: %w", route.Namespace, route.Name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| func (c *clientWrapper) UpdateBackendTLSPolicyStatus(ctx context.Context, policy ktypes.NamespacedName, status gatev1alpha2.PolicyStatus) error {
 | |
| 	if !c.isWatchedNamespace(policy.Namespace) {
 | |
| 		return fmt.Errorf("updating BackendTLSPolicy status %s/%s: namespace is not within watched namespaces", policy.Namespace, policy.Name)
 | |
| 	}
 | |
| 
 | |
| 	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 | |
| 		currentPolicy, err := c.factoriesGateway[c.lookupNamespace(policy.Namespace)].Gateway().V1alpha3().BackendTLSPolicies().Lister().BackendTLSPolicies(policy.Namespace).Get(policy.Name)
 | |
| 		if err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		ancestorStatuses := make([]gatev1alpha2.PolicyAncestorStatus, len(status.Ancestors))
 | |
| 		copy(ancestorStatuses, status.Ancestors)
 | |
| 
 | |
| 		// keep statuses added by other gateway controllers,
 | |
| 		// and statuses for Traefik gateway controller but not for the same Gateway as the one in parameter (AncestorRef).
 | |
| 		for _, ancestorStatus := range currentPolicy.Status.Ancestors {
 | |
| 			if ancestorStatus.ControllerName != controllerName {
 | |
| 				ancestorStatuses = append(ancestorStatuses, ancestorStatus)
 | |
| 				continue
 | |
| 			}
 | |
| 
 | |
| 			if slices.ContainsFunc(status.Ancestors, func(status gatev1alpha2.PolicyAncestorStatus) bool {
 | |
| 				return reflect.DeepEqual(ancestorStatus.AncestorRef, status.AncestorRef)
 | |
| 			}) {
 | |
| 				continue
 | |
| 			}
 | |
| 
 | |
| 			ancestorStatuses = append(ancestorStatuses, ancestorStatus)
 | |
| 		}
 | |
| 
 | |
| 		if len(ancestorStatuses) > 16 {
 | |
| 			return fmt.Errorf("failed to update BackendTLSPolicy %s/%s status: PolicyAncestor statuses count exceeds 16", policy.Namespace, policy.Name)
 | |
| 		}
 | |
| 
 | |
| 		// do not update status when nothing has changed.
 | |
| 		if policyAncestorStatusesEqual(currentPolicy.Status.Ancestors, ancestorStatuses) {
 | |
| 			return nil
 | |
| 		}
 | |
| 
 | |
| 		currentPolicy = currentPolicy.DeepCopy()
 | |
| 		currentPolicy.Status = gatev1alpha2.PolicyStatus{
 | |
| 			Ancestors: ancestorStatuses,
 | |
| 		}
 | |
| 
 | |
| 		if _, err = c.csGateway.GatewayV1alpha3().BackendTLSPolicies(policy.Namespace).UpdateStatus(ctx, currentPolicy, metav1.UpdateOptions{}); err != nil {
 | |
| 			// We have to return err itself here (not wrapped inside another error)
 | |
| 			// so that RetryOnConflict can identify it correctly.
 | |
| 			return err
 | |
| 		}
 | |
| 
 | |
| 		return nil
 | |
| 	})
 | |
| 	if err != nil {
 | |
| 		return fmt.Errorf("failed to update BackendTLSPolicy %s/%s status: %w", policy.Namespace, policy.Name, err)
 | |
| 	}
 | |
| 
 | |
| 	return nil
 | |
| }
 | |
| 
 | |
| // GetService returns the named service from the given namespace.
 | |
| func (c *clientWrapper) GetService(namespace, name string) (*corev1.Service, bool, error) {
 | |
| 	if !c.isWatchedNamespace(namespace) {
 | |
| 		return nil, false, fmt.Errorf("failed to get service %s/%s: namespace is not within watched namespaces", namespace, name)
 | |
| 	}
 | |
| 
 | |
| 	service, err := c.factoriesKube[c.lookupNamespace(namespace)].Core().V1().Services().Lister().Services(namespace).Get(name)
 | |
| 	exist, err := translateNotFoundError(err)
 | |
| 
 | |
| 	return service, exist, err
 | |
| }
 | |
| 
 | |
| // ListEndpointSlicesForService returns the EndpointSlices for the given service name in the given namespace.
 | |
| func (c *clientWrapper) ListEndpointSlicesForService(namespace, serviceName string) ([]*discoveryv1.EndpointSlice, error) {
 | |
| 	if !c.isWatchedNamespace(namespace) {
 | |
| 		return nil, fmt.Errorf("failed to get endpointslices for service %s/%s: namespace is not within watched namespaces", namespace, serviceName)
 | |
| 	}
 | |
| 
 | |
| 	serviceLabelRequirement, err := labels.NewRequirement(discoveryv1.LabelServiceName, selection.Equals, []string{serviceName})
 | |
| 	if err != nil {
 | |
| 		return nil, fmt.Errorf("failed to create service label selector requirement: %w", err)
 | |
| 	}
 | |
| 	serviceSelector := labels.NewSelector()
 | |
| 	serviceSelector = serviceSelector.Add(*serviceLabelRequirement)
 | |
| 
 | |
| 	return c.factoriesKube[c.lookupNamespace(namespace)].Discovery().V1().EndpointSlices().Lister().EndpointSlices(namespace).List(serviceSelector)
 | |
| }
 | |
| 
 | |
| // ListBackendTLSPoliciesForService returns the BackendTLSPolicy for the given service name in the given namespace.
 | |
| func (c *clientWrapper) ListBackendTLSPoliciesForService(namespace, serviceName string) ([]*gatev1alpha3.BackendTLSPolicy, error) {
 | |
| 	if !c.isWatchedNamespace(namespace) {
 | |
| 		return nil, fmt.Errorf("failed to get BackendTLSPolicies for service %s/%s: namespace is not within watched namespaces", namespace, serviceName)
 | |
| 	}
 | |
| 
 | |
| 	policies, err := c.factoriesGateway[c.lookupNamespace(namespace)].Gateway().V1alpha3().BackendTLSPolicies().Lister().BackendTLSPolicies(namespace).List(labels.Everything())
 | |
| 	if err != nil {
 | |
| 		return nil, fmt.Errorf("failed to list BackendTLSPolicies in namespace %s", namespace)
 | |
| 	}
 | |
| 
 | |
| 	var servicePolicies []*gatev1alpha3.BackendTLSPolicy
 | |
| 	for _, policy := range policies {
 | |
| 		for _, ref := range policy.Spec.TargetRefs {
 | |
| 			// The policy does not target the service.
 | |
| 			if (ref.Group != "" && ref.Group != groupCore) || ref.Kind != kindService || string(ref.Name) != serviceName {
 | |
| 				continue
 | |
| 			}
 | |
| 
 | |
| 			servicePolicies = append(servicePolicies, policy)
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	return servicePolicies, nil
 | |
| }
 | |
| 
 | |
| // GetSecret returns the named secret from the given namespace.
 | |
| func (c *clientWrapper) GetSecret(namespace, name string) (*corev1.Secret, bool, error) {
 | |
| 	if !c.isWatchedNamespace(namespace) {
 | |
| 		return nil, false, fmt.Errorf("failed to get secret %s/%s: namespace is not within watched namespaces", namespace, name)
 | |
| 	}
 | |
| 
 | |
| 	secret, err := c.factoriesSecret[c.lookupNamespace(namespace)].Core().V1().Secrets().Lister().Secrets(namespace).Get(name)
 | |
| 	exist, err := translateNotFoundError(err)
 | |
| 
 | |
| 	return secret, exist, err
 | |
| }
 | |
| 
 | |
| // GetConfigMap returns the named configMap from the given namespace.
 | |
| func (c *clientWrapper) GetConfigMap(namespace, name string) (*corev1.ConfigMap, bool, error) {
 | |
| 	if !c.isWatchedNamespace(namespace) {
 | |
| 		return nil, false, fmt.Errorf("failed to get configMap %s/%s: namespace is not within watched namespaces", namespace, name)
 | |
| 	}
 | |
| 
 | |
| 	configMap, err := c.factoriesKube[c.lookupNamespace(namespace)].Core().V1().ConfigMaps().Lister().ConfigMaps(namespace).Get(name)
 | |
| 	exist, err := translateNotFoundError(err)
 | |
| 
 | |
| 	return configMap, exist, err
 | |
| }
 | |
| 
 | |
| // lookupNamespace returns the lookup namespace key for the given namespace.
 | |
| // When listening on all namespaces, it returns the client-go identifier ("")
 | |
| // for all-namespaces. Otherwise, it returns the given namespace.
 | |
| // The distinction is necessary because we index all informers on the special
 | |
| // identifier iff all-namespaces are requested but receive specific namespace
 | |
| // identifiers from the Kubernetes API, so we have to bridge this gap.
 | |
| func (c *clientWrapper) lookupNamespace(namespace string) string {
 | |
| 	if c.isNamespaceAll {
 | |
| 		return metav1.NamespaceAll
 | |
| 	}
 | |
| 	return namespace
 | |
| }
 | |
| 
 | |
| // isWatchedNamespace checks to ensure that the namespace is being watched before we request
 | |
| // it to ensure we don't panic by requesting an out-of-watch object.
 | |
| func (c *clientWrapper) isWatchedNamespace(namespace string) bool {
 | |
| 	if c.isNamespaceAll {
 | |
| 		return true
 | |
| 	}
 | |
| 
 | |
| 	return slices.Contains(c.watchedNamespaces, namespace)
 | |
| }
 | |
| 
 | |
| // eventHandlerFunc will pass the obj on to the events channel or drop it.
 | |
| // This is so passing the events along won't block in the case of high volume.
 | |
| // The events are only used for signaling anyway so dropping a few is ok.
 | |
| func eventHandlerFunc(events chan<- interface{}, obj interface{}) {
 | |
| 	select {
 | |
| 	case events <- obj:
 | |
| 	default:
 | |
| 	}
 | |
| }
 | |
| 
 | |
| // translateNotFoundError will translate a "not found" error to a boolean return
 | |
| // value which indicates if the resource exists and a nil error.
 | |
| func translateNotFoundError(err error) (bool, error) {
 | |
| 	if kerror.IsNotFound(err) {
 | |
| 		return false, nil
 | |
| 	}
 | |
| 	return err == nil, err
 | |
| }
 | |
| 
 | |
| func gatewayStatusEqual(statusA, statusB gatev1.GatewayStatus) bool {
 | |
| 	return reflect.DeepEqual(statusA.Addresses, statusB.Addresses) &&
 | |
| 		listenersStatusEqual(statusA.Listeners, statusB.Listeners) &&
 | |
| 		conditionsEqual(statusA.Conditions, statusB.Conditions)
 | |
| }
 | |
| 
 | |
| func policyAncestorStatusesEqual(policyAncestorStatusesA, policyAncestorStatusesB []gatev1alpha2.PolicyAncestorStatus) bool {
 | |
| 	if len(policyAncestorStatusesA) != len(policyAncestorStatusesB) {
 | |
| 		return false
 | |
| 	}
 | |
| 
 | |
| 	for _, sA := range policyAncestorStatusesA {
 | |
| 		if !slices.ContainsFunc(policyAncestorStatusesB, func(sB gatev1alpha2.PolicyAncestorStatus) bool {
 | |
| 			return policyAncestorStatusEqual(sB, sA)
 | |
| 		}) {
 | |
| 			return false
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	for _, sB := range policyAncestorStatusesB {
 | |
| 		if !slices.ContainsFunc(policyAncestorStatusesA, func(sA gatev1alpha2.PolicyAncestorStatus) bool {
 | |
| 			return policyAncestorStatusEqual(sA, sB)
 | |
| 		}) {
 | |
| 			return false
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	return true
 | |
| }
 | |
| 
 | |
| func policyAncestorStatusEqual(sA, sB gatev1alpha2.PolicyAncestorStatus) bool {
 | |
| 	return sA.ControllerName == sB.ControllerName &&
 | |
| 		reflect.DeepEqual(sA.AncestorRef, sB.AncestorRef) &&
 | |
| 		conditionsEqual(sA.Conditions, sB.Conditions)
 | |
| }
 | |
| 
 | |
| func routeParentStatusesEqual(routeParentStatusesA, routeParentStatusesB []gatev1alpha2.RouteParentStatus) bool {
 | |
| 	if len(routeParentStatusesA) != len(routeParentStatusesB) {
 | |
| 		return false
 | |
| 	}
 | |
| 
 | |
| 	for _, sA := range routeParentStatusesA {
 | |
| 		if !slices.ContainsFunc(routeParentStatusesB, func(sB gatev1alpha2.RouteParentStatus) bool {
 | |
| 			return routeParentStatusEqual(sB, sA)
 | |
| 		}) {
 | |
| 			return false
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	for _, sB := range routeParentStatusesB {
 | |
| 		if !slices.ContainsFunc(routeParentStatusesA, func(sA gatev1alpha2.RouteParentStatus) bool {
 | |
| 			return routeParentStatusEqual(sA, sB)
 | |
| 		}) {
 | |
| 			return false
 | |
| 		}
 | |
| 	}
 | |
| 
 | |
| 	return true
 | |
| }
 | |
| 
 | |
| func routeParentStatusEqual(sA, sB gatev1alpha2.RouteParentStatus) bool {
 | |
| 	return sA.ControllerName == sB.ControllerName &&
 | |
| 		reflect.DeepEqual(sA.ParentRef, sB.ParentRef) &&
 | |
| 		conditionsEqual(sA.Conditions, sB.Conditions)
 | |
| }
 | |
| 
 | |
| func listenersStatusEqual(listenerA, listenerB []gatev1.ListenerStatus) bool {
 | |
| 	return slices.EqualFunc(listenerA, listenerB, func(lA gatev1.ListenerStatus, lB gatev1.ListenerStatus) bool {
 | |
| 		return lA.Name == lB.Name &&
 | |
| 			lA.AttachedRoutes == lB.AttachedRoutes &&
 | |
| 			conditionsEqual(lA.Conditions, lB.Conditions)
 | |
| 	})
 | |
| }
 | |
| 
 | |
| func conditionsEqual(conditionsA, conditionsB []metav1.Condition) bool {
 | |
| 	return slices.EqualFunc(conditionsA, conditionsB, func(cA metav1.Condition, cB metav1.Condition) bool {
 | |
| 		return cA.Type == cB.Type &&
 | |
| 			cA.Reason == cB.Reason &&
 | |
| 			cA.Status == cB.Status &&
 | |
| 			cA.Message == cB.Message &&
 | |
| 			cA.ObservedGeneration == cB.ObservedGeneration
 | |
| 	})
 | |
| }
 |