NO-JIRA: test(e2e/v2): sweep AGENTS.md conformance violations

test/e2e/v2/tests/control_plane_infrastructure_test.go

@@ -91,6 +91,7 @@ func InfrastructureRegistryValidationTest(getTestCtx internal.TestContextGetter)
 			testCtx := getTestCtx()
 
 			var podsNotBelongingToWorkloads []string
+			namespacesChecked := 0
 
 			for _, namespace := range internal.GetInfrastructureNamespaces() {
 				ns := &corev1.Namespace{}
@@ -99,6 +100,7 @@ func InfrastructureRegistryValidationTest(getTestCtx internal.TestContextGetter)
 					continue
 				}
 				Expect(err).NotTo(HaveOccurred(), "failed to get namespace %s", namespace)
+				namespacesChecked++
 
 				podList := &corev1.PodList{}
 				err = testCtx.MgmtClient.List(testCtx.Context, podList, &crclient.ListOptions{
@@ -122,6 +124,8 @@ func InfrastructureRegistryValidationTest(getTestCtx internal.TestContextGetter)
 				}
 			}
 
+			Expect(namespacesChecked).NotTo(BeZero(),
+				"expected at least one infrastructure namespace to exist")
 			Expect(podsNotBelongingToWorkloads).To(BeEmpty(),
 				"The following pods do not belong to any predefined infrastructure workload:\n%s",
 				strings.Join(podsNotBelongingToWorkloads, "\n"))
@@ -158,6 +162,10 @@ func InfrastructureResourceRequestsTest(getTestCtx internal.TestContextGetter) {
 						}
 					}
 
+					if len(matchingPods) == 0 {
+						Skip(fmt.Sprintf("no matching pods found for workload %s", workload.Name))
+					}
+
 					var failures []string
 					for _, pod := range matchingPods {
 						failures = append(failures, validateContainerResourceRequests(pod.Namespace, pod.Name, pod.Spec.Containers)...)

test/e2e/v2/tests/control_plane_workloads_test.go

@@ -82,7 +82,7 @@ func DeploymentGenerationTest(getTestCtx internal.TestContextGetter) {
 					}
 
 					deployment := &appsv1.Deployment{}
-					err := testCtx.MgmtClient.Get(context.Background(), crclient.ObjectKey{
+					err := testCtx.MgmtClient.Get(testCtx.Context, crclient.ObjectKey{
 						Namespace: testCtx.ControlPlaneNamespace,
 						Name:      workload.Name,
 					}, deployment)
@@ -722,6 +722,8 @@ func WorkloadRegistryValidationTest(getTestCtx internal.TestContextGetter) {
 				Namespace: testCtx.ControlPlaneNamespace,
 			})
 			Expect(err).NotTo(HaveOccurred(), "failed to list pods in control plane namespace")
+			Expect(podList.Items).NotTo(BeEmpty(),
+				"expected pods in control plane namespace %s", testCtx.ControlPlaneNamespace)
 
 			workloadSelectors := make(map[string]labels.Selector)
 			for _, workload := range workloads {
@@ -767,7 +769,7 @@ func SecurityContextUIDTest(getTestCtx internal.TestContextGetter) {
 
 			// Get the control plane namespace to check for UID annotation
 			controlPlaneNamespace := &corev1.Namespace{}
-			err := testCtx.MgmtClient.Get(context.Background(), crclient.ObjectKey{Name: testCtx.ControlPlaneNamespace}, controlPlaneNamespace)
+			err := testCtx.MgmtClient.Get(testCtx.Context, crclient.ObjectKey{Name: testCtx.ControlPlaneNamespace}, controlPlaneNamespace)
 			Expect(err).NotTo(HaveOccurred(), "failed to get namespace %s", testCtx.ControlPlaneNamespace)
 
 			uid, ok := controlPlaneNamespace.Annotations["hypershift.openshift.io/default-security-context-uid"]

test/e2e/v2/tests/etcd_chaos_test.go

@@ -126,13 +126,14 @@ func EtcdKillRandomMembersTest(getTestCtx internal.TestContextGetter) {
 		ctx := testCtx.Context
 		cpNamespace := testCtx.ControlPlaneNamespace
 
-		guestClient := testCtx.GetHostedClusterClient()
+		hcClient := testCtx.GetHostedClusterClient()
 
 		// Create marker data that should survive the chaos
-		markerCM := createMarkerConfigMap(ctx, guestClient)
+		markerCM := createMarkerConfigMap(ctx, hcClient)
 		DeferCleanup(func() {
-			if err := guestClient.Delete(ctx, markerCM); err != nil && !apierrors.IsNotFound(err) {
-				GinkgoWriter.Printf("Warning: failed to cleanup marker ConfigMap: %v\n", err)
+			err := hcClient.Delete(ctx, markerCM)
+			if err != nil && !apierrors.IsNotFound(err) {
+				Expect(err).NotTo(HaveOccurred(), "cleanup: failed to delete marker ConfigMap %s", markerCM.Name)
 			}
 		})
 
@@ -160,7 +161,7 @@ func EtcdKillRandomMembersTest(getTestCtx internal.TestContextGetter) {
 
 		waitForEtcdConvergence(ctx, testCtx.MgmtClient, cpNamespace, ptr.Deref(etcdSts.Spec.Replicas, 0))
 
-		verifyMarkerSurvived(ctx, guestClient, markerCM)
+		verifyMarkerSurvived(ctx, hcClient, markerCM)
 	})
 }
 
@@ -173,13 +174,14 @@ func EtcdKillAllMembersTest(getTestCtx internal.TestContextGetter) {
 		ctx := testCtx.Context
 		cpNamespace := testCtx.ControlPlaneNamespace
 
-		guestClient := testCtx.GetHostedClusterClient()
+		hcClient := testCtx.GetHostedClusterClient()
 
 		// Create marker data that should survive the chaos
-		markerCM := createMarkerConfigMap(ctx, guestClient)
+		markerCM := createMarkerConfigMap(ctx, hcClient)
 		DeferCleanup(func() {
-			if err := guestClient.Delete(ctx, markerCM); err != nil && !apierrors.IsNotFound(err) {
-				GinkgoWriter.Printf("Warning: failed to cleanup marker ConfigMap: %v\n", err)
+			err := hcClient.Delete(ctx, markerCM)
+			if err != nil && !apierrors.IsNotFound(err) {
+				Expect(err).NotTo(HaveOccurred(), "cleanup: failed to delete marker ConfigMap %s", markerCM.Name)
 			}
 		})
 
@@ -236,7 +238,7 @@ func EtcdKillAllMembersTest(getTestCtx internal.TestContextGetter) {
 
 		waitForEtcdConvergence(ctx, testCtx.MgmtClient, cpNamespace, ptr.Deref(etcdSts.Spec.Replicas, 0))
 
-		verifyMarkerSurvived(ctx, guestClient, markerCM)
+		verifyMarkerSurvived(ctx, hcClient, markerCM)
 	})
 }
 

test/e2e/v2/tests/hosted_cluster_aws_test.go

@@ -56,6 +56,10 @@ func EnsureDefaultSecurityGroupTagsTest(getTestCtx internal.TestContextGetter) {
 				Skip("default security group tags test is only for AWS platform")
 			}
 
+			Expect(hc.Status.Platform).NotTo(BeNil(),
+				"HostedCluster %s/%s should have platform status", hc.Namespace, hc.Name)
+			Expect(hc.Status.Platform.AWS).NotTo(BeNil(),
+				"HostedCluster %s/%s should have AWS platform status", hc.Namespace, hc.Name)
 			sgID := hc.Status.Platform.AWS.DefaultWorkerSecurityGroupID
 			Expect(sgID).NotTo(BeEmpty(), "HostedCluster status should have DefaultWorkerSecurityGroupID set")
 
@@ -202,6 +206,12 @@ func AWSCCMWithCustomizationsTest(getTestCtx internal.TestContextGetter) {
 					},
 				}
 				Expect(hcClient.Create(tc.Context, testSvc)).To(Succeed(), "failed to create LoadBalancer service")
+				DeferCleanup(func() {
+					err := hcClient.Delete(tc.Context, testSvc)
+					if err != nil && !apierrors.IsNotFound(err) {
+						Expect(err).NotTo(HaveOccurred(), "cleanup: failed to delete test service %s", testSvc.Name)
+					}
+				})
 
 				var lbHostname string
 				Eventually(func(g Gomega) {

test/e2e/v2/tests/hosted_cluster_ccm_test.go

@@ -17,7 +17,6 @@ limitations under the License.
 package tests
 
 import (
-	"context"
 	"strings"
 
 	. "github.com/onsi/ginkgo/v2"
@@ -51,7 +50,7 @@ func GCPCloudControllerManagerTest(getTestCtx internal.TestContextGetter) {
 				Expect(hostedClusterClient).NotTo(BeNil(), "hosted cluster client is nil; HostedCluster may not have KubeConfig status set")
 
 				nodes = &corev1.NodeList{}
-				Expect(hostedClusterClient.List(context.Background(), nodes)).To(Succeed())
+				Expect(hostedClusterClient.List(testCtx.Context, nodes)).To(Succeed())
 				Expect(nodes.Items).NotTo(BeEmpty(), "cluster should have nodes")
 			})
 
@@ -112,6 +111,7 @@ var _ = Describe("Hosted Cluster CCM", Label("hosted-cluster-ccm"), func() {
 	BeforeEach(func() {
 		testCtx = internal.GetTestContext()
 		Expect(testCtx).NotTo(BeNil(), "test context should be set up in BeforeSuite")
+		testCtx.ValidateHostedCluster()
 	})
 
 	RegisterHostedClusterCCMTests(func() *internal.TestContext { return testCtx })

test/e2e/v2/tests/hosted_cluster_health_test.go

@@ -80,6 +80,8 @@ func EnsureCAPIFinalizersTest(getTestCtx internal.TestContextGetter) {
 	When("CAPI components are deployed", func() {
 		It("should have component finalizers on all CAPI deployments", func() {
 			tc := getTestCtx()
+			Expect(hcc.CAPIComponents).NotTo(BeEmpty(),
+				"expected CAPI components to be defined in HostedControlPlaneConfiguration")
 			for _, name := range hcc.CAPIComponents {
 				deployment := &appsv1.Deployment{}
 				Expect(tc.MgmtClient.Get(tc.Context, crclient.ObjectKey{

test/e2e/v2/tests/hosted_cluster_image_registry_test.go

@@ -144,6 +144,8 @@ func ImageRegistryCapabilityEnabledTest(getTestCtx internal.TestContextGetter) {
 				}
 				Expect(hc.Spec.Platform.GCP).NotTo(BeNil(),
 					"GCP platform spec must be set for GCP HostedCluster %s/%s", hc.Namespace, hc.Name)
+				Expect(hc.Spec.Platform.GCP.WorkloadIdentity).NotTo(BeNil(),
+					"GCP platform spec should have WorkloadIdentity configured for HostedCluster %s/%s", hc.Namespace, hc.Name)
 				imageRegistryEmail = string(hc.Spec.Platform.GCP.WorkloadIdentity.ServiceAccountsEmails.ImageRegistry)
 				Expect(imageRegistryEmail).NotTo(BeEmpty(),
 					"imageRegistry service account email must be set in GCP WorkloadIdentity config for %s/%s",

test/e2e/v2/tests/hosted_cluster_metrics_test.go

@@ -189,7 +189,7 @@ func EnsureMetricsForwarderWorkingTest(getTestCtx internal.TestContextGetter) {
 				g.Expect(pod.Status.Phase).To(Equal(corev1.PodRunning), "prometheus pod should be running")
 			}, 5*time.Minute, 10*time.Second).Should(Succeed())
 
-			By("Verifying guest Prometheus is scraping kube-apiserver via the metrics-forwarder")
+			By("Verifying hosted cluster Prometheus is scraping kube-apiserver via the metrics-forwarder")
 			Eventually(func(g Gomega) {
 				output, err := v2util.RunCommandInPod(tc.Context, hcClientset, hcRestConfig,
 					monitoringNamespace, promPodName, "prometheus",

test/e2e/v2/tests/hosted_cluster_security_test.go

@@ -45,12 +45,12 @@ import (
 )
 
 func RegisterHostedClusterSecurityTests(getTestCtx internal.TestContextGetter) {
-	EnsureGuestWebhooksValidatedTest(getTestCtx)
+	EnsureHostedClusterWebhooksValidatedTest(getTestCtx)
 	EnsureAdmissionPoliciesTest(getTestCtx)
 	EnsureNetworkPoliciesTest(getTestCtx)
 }
 
-func EnsureGuestWebhooksValidatedTest(getTestCtx internal.TestContextGetter) {
+func EnsureHostedClusterWebhooksValidatedTest(getTestCtx internal.TestContextGetter) {
 	When("a webhook targeting a control plane service is created in the hosted cluster", func() {
 		It("should be automatically deleted", func() {
 			tc := getTestCtx()

test/e2e/v2/tests/nodepool_autoscaling_test.go

@@ -47,7 +47,7 @@ func AutoscalingScaleUpDownTest(getTestCtx internal.TestContextGetter) {
 		testCtx.ValidateHostedClusterClient()
 
 		hc := testCtx.GetHostedCluster()
-		guestClient := testCtx.GetHostedClusterClient()
+		hcClient := testCtx.GetHostedClusterClient()
 		ctx := testCtx.Context
 
 		// Find the default NodePool to copy platform config
@@ -63,14 +63,16 @@ func AutoscalingScaleUpDownTest(getTestCtx internal.TestContextGetter) {
 		GinkgoWriter.Printf("Created autoscaling NodePool %s with min=1, max=3\n", autoscalingNP.Name)
 
 		// Ensure cleanup
-		defer cleanupNodePool(ctx, testCtx.MgmtClient, autoscalingNP)
+		DeferCleanup(func() {
+			cleanupNodePool(ctx, testCtx.MgmtClient, autoscalingNP)
+		})
 
 		npLabelSelector := e2eutil.WithClientOptions(crclient.MatchingLabelsSelector{
 			Selector: labels.SelectorFromSet(labels.Set{hyperv1.NodePoolLabel: autoscalingNP.Name}),
 		})
 
 		// Wait for NodePool to be ready with 1 node (min replicas)
-		nodes := e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, guestClient, 1, hc.Spec.Platform.Type, fmt.Sprintf("for NodePool %s", autoscalingNP.Name), npLabelSelector)
+		nodes := e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, hcClient, 1, hc.Spec.Platform.Type, fmt.Sprintf("for NodePool %s", autoscalingNP.Name), npLabelSelector)
 		Expect(nodes).To(HaveLen(1), "should have exactly 1 node initially")
 
 		// Get node capacity for workload sizing
@@ -83,19 +85,21 @@ func AutoscalingScaleUpDownTest(getTestCtx internal.TestContextGetter) {
 		// cluster autoscaler must scale it up.
 		workloadMemRequest := *resource.NewQuantity(bytes/2, resource.BinarySI)
 		workload := newAutoscalingWorkload(3, workloadMemRequest, autoscalingLabel)
-		err = guestClient.Create(ctx, workload)
+		err = hcClient.Create(ctx, workload)
 		Expect(err).NotTo(HaveOccurred(), "failed to create workload")
 
-		defer cleanupWorkload(ctx, guestClient, workload)
+		DeferCleanup(func() {
+			cleanupWorkload(ctx, hcClient, workload)
+		})
 
 		// Wait for scale-up to 3 nodes
-		e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, guestClient, 3, hc.Spec.Platform.Type, fmt.Sprintf("for NodePool %s", autoscalingNP.Name), npLabelSelector)
+		e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, hcClient, 3, hc.Spec.Platform.Type, fmt.Sprintf("for NodePool %s", autoscalingNP.Name), npLabelSelector)
 
 		// Delete workload to trigger scale-down
-		cleanupWorkload(ctx, guestClient, workload)
+		cleanupWorkload(ctx, hcClient, workload)
 
 		// Wait for scale-down to 1 node (min replicas)
-		e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, guestClient, 1, hc.Spec.Platform.Type, fmt.Sprintf("for NodePool %s", autoscalingNP.Name), npLabelSelector)
+		e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, hcClient, 1, hc.Spec.Platform.Type, fmt.Sprintf("for NodePool %s", autoscalingNP.Name), npLabelSelector)
 	})
 }
 
@@ -110,7 +114,7 @@ func AutoscalingBalancingTest(getTestCtx internal.TestContextGetter) {
 		e2eutil.GinkgoAtLeast(e2eutil.Version420)
 
 		hc := testCtx.GetHostedCluster()
-		guestClient := testCtx.GetHostedClusterClient()
+		hcClient := testCtx.GetHostedClusterClient()
 		ctx := testCtx.Context
 		cpNamespace := testCtx.ControlPlaneNamespace
 
@@ -133,15 +137,12 @@ func AutoscalingBalancingTest(getTestCtx internal.TestContextGetter) {
 
 		DeferCleanup(func() {
 			latest := &hyperv1.HostedCluster{}
-			if err := testCtx.MgmtClient.Get(ctx, crclient.ObjectKeyFromObject(hc), latest); err != nil {
-				GinkgoWriter.Printf("Warning: failed to get HostedCluster for cleanup: %v\n", err)
-				return
-			}
+			Expect(testCtx.MgmtClient.Get(ctx, crclient.ObjectKeyFromObject(hc), latest)).To(Succeed(),
+				"cleanup: failed to get HostedCluster for autoscaling config reset")
 			patch := crclient.MergeFrom(latest.DeepCopy())
 			latest.Spec.Autoscaling = hyperv1.ClusterAutoscaling{}
-			if err := testCtx.MgmtClient.Patch(ctx, latest, patch); err != nil {
-				GinkgoWriter.Printf("Warning: failed to reset autoscaler config: %v\n", err)
-			}
+			Expect(testCtx.MgmtClient.Patch(ctx, latest, patch)).To(Succeed(),
+				"cleanup: failed to reset autoscaler config on HostedCluster")
 		})
 
 		// Wait for autoscaler deployment to pick up the new config
@@ -184,12 +185,16 @@ func AutoscalingBalancingTest(getTestCtx internal.TestContextGetter) {
 		autoscalingNP1 := buildAutoscalingNodePool(defaultNP, 1, 3, np1Labels)
 		err = testCtx.MgmtClient.Create(ctx, autoscalingNP1)
 		Expect(err).NotTo(HaveOccurred(), "failed to create first autoscaling NodePool")
-		defer cleanupNodePool(ctx, testCtx.MgmtClient, autoscalingNP1)
+		DeferCleanup(func() {
+			cleanupNodePool(ctx, testCtx.MgmtClient, autoscalingNP1)
+		})
 
 		autoscalingNP2 := buildAutoscalingNodePool(defaultNP, 1, 3, np2Labels)
 		err = testCtx.MgmtClient.Create(ctx, autoscalingNP2)
 		Expect(err).NotTo(HaveOccurred(), "failed to create second autoscaling NodePool")
-		defer cleanupNodePool(ctx, testCtx.MgmtClient, autoscalingNP2)
+		DeferCleanup(func() {
+			cleanupNodePool(ctx, testCtx.MgmtClient, autoscalingNP2)
+		})
 
 		np1LabelSelector := e2eutil.WithClientOptions(crclient.MatchingLabelsSelector{
 			Selector: labels.SelectorFromSet(labels.Set{hyperv1.NodePoolLabel: autoscalingNP1.Name}),
@@ -199,8 +204,8 @@ func AutoscalingBalancingTest(getTestCtx internal.TestContextGetter) {
 		})
 
 		// Wait for initial nodes (1 per NodePool at min replicas)
-		nodes := e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, guestClient, 1, hc.Spec.Platform.Type, "for NP1", np1LabelSelector)
-		e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, guestClient, 1, hc.Spec.Platform.Type, "for NP2", np2LabelSelector)
+		nodes := e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, hcClient, 1, hc.Spec.Platform.Type, "for NP1", np1LabelSelector)
+		e2eutil.WaitForNReadyNodesWithOptions(GinkgoTB(), ctx, hcClient, 1, hc.Spec.Platform.Type, "for NP2", np2LabelSelector)
 
 		// Get node capacity for workload sizing
 		memCapacity := nodes[0].Status.Allocatable[corev1.ResourceMemory]
@@ -210,9 +215,11 @@ func AutoscalingBalancingTest(getTestCtx internal.TestContextGetter) {
 		// Create workload targeting the autoscaling NodePools via the shared label.
 		workloadMemRequest := *resource.NewQuantity(bytes/2, resource.BinarySI)
 		workload := newAutoscalingWorkload(4, workloadMemRequest, sharedLabel)
-		err = guestClient.Create(ctx, workload)
+		err = hcClient.Create(ctx, workload)
 		Expect(err).NotTo(HaveOccurred(), "failed to create workload")
-		defer cleanupWorkload(ctx, guestClient, workload)
+		DeferCleanup(func() {
+			cleanupWorkload(ctx, hcClient, workload)
+		})
 
 		// Wait for total 4 nodes across both NPs, then verify balanced distribution
 		Eventually(func() (bool, error) {