add operator e2e suites

Signed-off-by: zhzhuang-zju <m17799853869@163.com>

.github/workflows/ci.yml

@@ -161,3 +161,59 @@ jobs:
         with:
           name: karmada_kind_log_${{ matrix.k8s }}
           path: /tmp/karmada/
+
+  e2e-operator:
+    name: operator e2e test
+    needs: build
+    runs-on: ubuntu-22.04
+    strategy:
+      fail-fast: false
+      matrix:
+        # Here support the latest three minor releases of Kubernetes, this can be considered to be roughly
+        # the same as the End of Life of the Kubernetes release: https://kubernetes.io/releases/
+        # Please remember to update the CI Schedule Workflow when we add a new version.
+        k8s: [ v1.29.0, v1.30.0, v1.31.0 ]
+    steps:
+      # Free up disk space on Ubuntu
+      - name: Free Disk Space (Ubuntu)
+        uses: jlumbroso/free-disk-space@main
+        with:
+          # this might remove tools that are actually needed, if set to "true" but frees about 6 GB
+          tool-cache: false
+          # all of these default to true, but feel free to set to "false" if necessary for your workflow
+          android: true
+          dotnet: true
+          haskell: true
+          large-packages: false
+          docker-images: false
+          swap-storage: false
+      - name: checkout code
+        uses: actions/checkout@v4
+        with:
+          # Number of commits to fetch. 0 indicates all history for all branches and tags.
+          # We need to guess version via git tags.
+          fetch-depth: 0
+      - name: install Go
+        uses: actions/setup-go@v5
+        with:
+          go-version-file: go.mod
+      - name: setup operator e2e test environment
+        run: |
+          export CLUSTER_VERSION=kindest/node:${{ matrix.k8s }}
+          hack/operator-e2e-environment.sh
+      - name: run e2e
+        run: |
+          export ARTIFACTS_PATH=${{ github.workspace }}/karmada-operator-e2e-logs/${{ matrix.k8s }}/
+          hack/run-e2e-operator.sh
+      - name: upload logs
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: karmada_operator_e2e_log_${{ matrix.k8s }}
+          path: ${{ github.workspace }}/karmada-operator-e2e-logs/${{ matrix.k8s }}/
+      - name: upload kind logs
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: karmada_operator_kind_log_${{ matrix.k8s }}
+          path: /tmp/karmada/

hack/operator-e2e-environment.sh

@@ -0,0 +1,82 @@
+#!/usr/bin/env bash
+# Copyright 2025 The Karmada Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+set -o errexit
+set -o nounset
+set -o pipefail
+
+# This script is used in workflow to set up a local karmada-operator e2e testing environment.
+# It deploys karmada-operator and related resources to the host cluster.
+# This script depends on utils in: ${REPO_ROOT}/hack/util.sh.
+
+function usage() {
+    echo "Usage:"
+    echo "    hack/operator-e2e-environment.sh [-h]"
+    echo "    h: print help information"
+}
+
+function getCrdsDir() {
+  local path=$1
+  local url=$2
+  local key=$(echo "$url" | xargs)  # Trim whitespace using xargs
+  local hash=$(echo -n "$key" | sha256sum | awk '{print $1}')  # Calculate SHA256 hash
+  local hashedKey=${hash:0:64}  # Take the first 64 characters of the hash
+  echo "${path}/cache/${hashedKey}"
+}
+
+while getopts 'h' OPT; do
+    case $OPT in
+        h)
+          usage
+          exit 0
+          ;;
+        ?)
+          usage
+          exit 1
+          ;;
+    esac
+done
+
+REPO_ROOT=$(dirname "${BASH_SOURCE[0]}")/..
+source "${REPO_ROOT}"/hack/util.sh
+KARMADA_SYSTEM_NAMESPACE="karmada-system"
+
+# variable define
+export KUBECONFIG_PATH=${KUBECONFIG_PATH:-"${HOME}/.kube"}
+export MAIN_KUBECONFIG=${MAIN_KUBECONFIG:-"${KUBECONFIG_PATH}/karmada.config"}
+export HOST_CLUSTER_NAME=${HOST_CLUSTER_NAME:-"karmada-host"}
+
+# step1. set up a base development environment
+"${REPO_ROOT}"/hack/setup-dev-base.sh
+export KUBECONFIG="${MAIN_KUBECONFIG}"
+
+# step2. deploy karmada-operator
+"${REPO_ROOT}"/hack/deploy-karmada-operator.sh "${MAIN_KUBECONFIG}" "${HOST_CLUSTER_NAME}"
+
+# step3. install karmada instance by karmada-operator
+# step3.1 prepare the local crds
+echo "Prepare the local crds"
+cd  ${REPO_ROOT}/charts/karmada/
+cp -r _crds crds
+tar -zcvf ../../crds.tar.gz crds
+cd -
+
+# step3.2 copy the local crds.tar.gz file to the specified path of the karmada-operator, so that the karmada-operator will skip the step of downloading CRDs.
+CRDTARBALL_URL="http://local"
+DATA_DIR="/var/lib/karmada"
+CRD_CACHE_DIR=$(getCrdsDir "${DATA_DIR}" "${CRDTARBALL_URL}")
+OPERATOR_POD_NAME=$(kubectl --kubeconfig="${MAIN_KUBECONFIG}" --context="${HOST_CLUSTER_NAME}" get pods -n ${KARMADA_SYSTEM_NAMESPACE} -l karmada-app=karmada-operator -o custom-columns=NAME:.metadata.name --no-headers)
+kubectl --kubeconfig="${MAIN_KUBECONFIG}" --context="${HOST_CLUSTER_NAME}" exec -i ${OPERATOR_POD_NAME} -n ${KARMADA_SYSTEM_NAMESPACE} -- mkdir -p ${CRD_CACHE_DIR}
+kubectl --kubeconfig="${MAIN_KUBECONFIG}" --context="${HOST_CLUSTER_NAME}" cp ${REPO_ROOT}/crds.tar.gz ${KARMADA_SYSTEM_NAMESPACE}/${OPERATOR_POD_NAME}:${CRD_CACHE_DIR}

hack/run-e2e-operator.sh

@@ -0,0 +1,63 @@
+#!/usr/bin/env bash
+# Copyright 2025 The Karmada Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+set -o errexit
+set -o nounset
+set -o pipefail
+
+# This script runs e2e test against on a kubernetes cluster which deployed karmada operator.
+# You should prepare your environment in advance, for example, use hack/operator-e2e-environment.sh to set up the environment.
+#
+# Usage: hack/run-e2e-operator.sh
+# Example 1: hack/run-e2e-operator.sh (run operator e2e with default config)
+
+KUBECONFIG_PATH=${KUBECONFIG_PATH:-"${HOME}/.kube"}
+HOST_KUBECONFIG=${HOST_KUBECONFIG:-"$KUBECONFIG_PATH/karmada.config"}
+
+# RUNNING_ON_KIND indicates if current testing against on a kind cluster.
+# Defaults to true.
+# For kind cluster, the kind related logs will be collected after the testing.
+RUNNING_ON_KIND=${RUNNING_ON_KIND:-true}
+
+HOST_CLUSTER_NAME=${HOST_CLUSTER_NAME:-"karmada-host"}
+
+ARTIFACTS_PATH=${ARTIFACTS_PATH:-"${HOME}/karmada-operator-e2e-logs"}
+mkdir -p "$ARTIFACTS_PATH"
+
+# Install ginkgo
+GO111MODULE=on go install github.com/onsi/ginkgo/v2/ginkgo
+
+# Run e2e
+export KUBECONFIG=${HOST_KUBECONFIG}
+
+set +e
+ginkgo -v --race --trace --fail-fast -p --randomize-all ./test/e2e/suites/operator
+TESTING_RESULT=$?
+
+# Collect logs
+echo "Collect logs to $ARTIFACTS_PATH..."
+cp "$HOST_KUBECONFIG" "$ARTIFACTS_PATH"
+
+if [ "$RUNNING_ON_KIND" = true ]; then
+  echo "Collecting $HOST_CLUSTER_NAME logs..."
+  mkdir -p "$ARTIFACTS_PATH/$HOST_CLUSTER_NAME"
+  kind export logs --name="$HOST_CLUSTER_NAME" "$ARTIFACTS_PATH/$HOST_CLUSTER_NAME"
+fi
+
+echo "Collected logs at $ARTIFACTS_PATH:"
+ls -al "$ARTIFACTS_PATH"
+
+exit $TESTING_RESULT

test/e2e/framework/cluster.go

@@ -264,7 +264,7 @@ func UpdateClusterLabels(client karmada.Interface, clusterName string, labels ma
 			return false, err
 		}
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // DeleteClusterLabels deletes cluster labels if it exists.
@@ -286,7 +286,7 @@ func DeleteClusterLabels(client karmada.Interface, clusterName string, labels ma
 			return false, err
 		}
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // GetClusterNamesFromClusters will get Clusters' names form Clusters Object.
@@ -306,7 +306,7 @@ func WaitClusterFitWith(c client.Client, clusterName string, fit func(cluster *c
 			return false, err
 		}
 		return fit(currentCluster), nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // LoadRESTClientConfig creates a rest.Config using the passed kubeconfig. If context is empty, current context in kubeconfig will be used.
@@ -363,5 +363,5 @@ func UpdateClusterStatusCondition(client karmada.Interface, clusterName string,
 			return false, err
 		}
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/clusterresourcebinding.go

@@ -34,5 +34,5 @@ func WaitClusterResourceBindingFitWith(client karmada.Interface, name string, fi
 			return false
 		}
 		return fit(clusterResourceBinding)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/configmap.go

@@ -68,7 +68,7 @@ func WaitConfigMapPresentOnClusterFitWith(cluster, namespace, name string, fit f
 			return false
 		}
 		return fit(configmap)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // UpdateConfigMapWithPatch update configmap with patch bytes.
@@ -99,7 +99,7 @@ func WaitConfigMapDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get configmap(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitConfigMapDisappearOnClusters wait configmap disappear on member clusters until timeout.

test/e2e/framework/constant.go

@@ -19,10 +19,10 @@ package framework
 import "time"
 
 const (
-	// pollInterval defines the interval time for a poll operation.
+	// PollInterval defines the interval time for a poll operation.
-	pollInterval = 5 * time.Second
+	PollInterval = 5 * time.Second
-	// pollTimeout defines the time after which the poll operation times out.
+	// PollTimeout defines the time after which the poll operation times out.
-	pollTimeout = 420 * time.Second
+	PollTimeout = 420 * time.Second
 	// metricsCreationDelay defines the maximum time metrics not yet available for pod.
 	metricsCreationDelay = 2 * time.Minute
 )

test/e2e/framework/customresourcedefine.go

@@ -74,7 +74,7 @@ func WaitCRDPresentOnClusters(client karmada.Interface, clusters []string, crdAP
 				cluster, err := FetchCluster(client, clusterName)
 				g.Expect(err).NotTo(gomega.HaveOccurred())
 				return helper.IsAPIEnabled(cluster.Status.APIEnablements, crdAPIVersion, crdKind), nil
-			}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+			}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 		}
 	})
 }
@@ -90,7 +90,7 @@ func WaitCRDDisappearedOnClusters(clusters []string, crdName string) {
 			gomega.Eventually(func() bool {
 				_, err := clusterDynamicClient.Resource(crdGVR).Get(context.TODO(), crdName, metav1.GetOptions{})
 				return apierrors.IsNotFound(err)
-			}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+			}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 		}
 	})
 }
@@ -108,5 +108,5 @@ func WaitCRDFitWith(client dynamic.Interface, crdName string, fit func(crd *apie
 			return false
 		}
 		return fit(crd)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/deployment.go

@@ -55,7 +55,7 @@ func UpdateDeploymentPaused(client kubernetes.Interface, deployment *appsv1.Depl
 			deploy.Spec.Paused = paused
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -79,7 +79,7 @@ func WaitDeploymentPresentOnClusterFitWith(cluster, namespace, name string, fit
 			return false
 		}
 		return fit(dep)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitDeploymentFitWith wait deployment sync with fit func.
@@ -90,7 +90,7 @@ func WaitDeploymentFitWith(client kubernetes.Interface, namespace, name string,
 			return false
 		}
 		return fit(dep)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitDeploymentPresentOnClustersFitWith wait deployment present on cluster sync with fit func.
@@ -111,7 +111,7 @@ func WaitDeploymentStatus(client kubernetes.Interface, deployment *appsv1.Deploy
 				return false
 			}
 			return CheckDeploymentReadyStatus(deploy, replicas)
-		}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+		}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 	})
 }
 
@@ -132,7 +132,7 @@ func WaitDeploymentDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get deployment(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitDeploymentDisappearOnClusters wait deployment disappear on member clusters until timeout.
@@ -155,7 +155,7 @@ func UpdateDeploymentReplicas(client kubernetes.Interface, deployment *appsv1.De
 			deploy.Spec.Replicas = &replicas
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -170,7 +170,7 @@ func UpdateDeploymentAnnotations(client kubernetes.Interface, deployment *appsv1
 			deploy.Annotations = annotations
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -190,7 +190,7 @@ func AppendDeploymentAnnotations(client kubernetes.Interface, deployment *appsv1
 			}
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -205,7 +205,7 @@ func UpdateDeploymentLabels(client kubernetes.Interface, deployment *appsv1.Depl
 			deploy.Labels = labels
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -220,7 +220,7 @@ func UpdateDeploymentVolumes(client kubernetes.Interface, deployment *appsv1.Dep
 			deploy.Spec.Template.Spec.Volumes = volumes
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -235,7 +235,7 @@ func UpdateDeploymentServiceAccountName(client kubernetes.Interface, deployment
 			deploy.Spec.Template.Spec.ServiceAccountName = serviceAccountName
 			_, err = client.AppsV1().Deployments(deploy.Namespace).Update(context.TODO(), deploy, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -252,7 +252,7 @@ func ExtractTargetClustersFrom(c client.Client, deployment *appsv1.Deployment) [
 			return false, nil
 		}
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 
 	targetClusterNames := make([]string, 0, len(binding.Spec.Clusters))
 	for _, cluster := range binding.Spec.Clusters {
@@ -282,7 +282,7 @@ func WaitDeploymentGetByClientFitWith(client kubernetes.Interface, namespace, na
 				return false
 			}
 			return fit(dep)
-		}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+		}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 	})
 }
 
@@ -305,6 +305,6 @@ func WaitDeploymentReplicasFitWith(clusters []string, namespace, name string, ex
 			}
 			klog.Infof("The total replicas of deployment(%s/%s) is %d", namespace, name, totalReplicas)
 			return totalReplicas == expectReplicas
-		}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+		}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 	})
 }

test/e2e/framework/events.go

@@ -35,5 +35,5 @@ func WaitEventFitWith(kubeClient kubernetes.Interface, namespace string, involve
 		})
 		gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
 		return slices.ContainsFunc(eventList.Items, fit)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/federatedresourcequota.go

@@ -86,6 +86,6 @@ func WaitFederatedResourceQuotaCollectStatus(client karmada.Interface, namespace
 			}
 
 			return true, nil
-		}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+		}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 	})
 }

test/e2e/framework/job.go

@@ -67,7 +67,7 @@ func WaitJobPresentOnClusterFitWith(cluster, namespace, name string, fit func(jo
 			return false
 		}
 		return fit(dep)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitJobPresentOnClustersFitWith wait job present on cluster sync with fit func.
@@ -96,7 +96,7 @@ func WaitJobDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get job(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitJobDisappearOnClusters wait job disappear on member clusters until timeout.

test/e2e/framework/mcs.go

@@ -48,7 +48,7 @@ func UpdateMultiClusterService(client karmada.Interface, mcs *networkingv1alpha1
 			mcsExist.Spec = mcs.Spec
 			_, err = client.NetworkingV1alpha1().MultiClusterServices(mcsExist.Namespace).Update(context.TODO(), mcsExist, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -68,5 +68,5 @@ func WaitMultiClusterServicePresentOnClustersFitWith(client karmada.Interface, n
 		gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
 
 		return fit(mcs)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/namespace.go

@@ -52,7 +52,7 @@ func WaitNamespacePresentOnClusterByClient(client kubernetes.Interface, name str
 		_, err := client.CoreV1().Namespaces().Get(context.TODO(), name, metav1.GetOptions{})
 		g.Expect(err).NotTo(gomega.HaveOccurred())
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitNamespacePresentOnCluster wait namespace present on cluster until timeout.
@@ -65,7 +65,7 @@ func WaitNamespacePresentOnCluster(cluster, name string) {
 		_, err := clusterClient.CoreV1().Namespaces().Get(context.TODO(), name, metav1.GetOptions{})
 		g.Expect(err).NotTo(gomega.HaveOccurred())
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitNamespacePresentOnClusters wait namespace present on clusters until timeout.
@@ -94,7 +94,7 @@ func WaitNamespaceDisappearOnCluster(cluster, name string) {
 
 		klog.Errorf("Failed to get namespace(%s) on cluster(%s), err: %v", name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitNamespaceDisappearOnClusters wait namespace disappear on clusters until timeout.
@@ -117,6 +117,6 @@ func UpdateNamespaceLabels(client kubernetes.Interface, namespace *corev1.Namesp
 			ns.Labels = labels
 			_, err = client.CoreV1().Namespaces().Update(context.TODO(), ns, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }

test/e2e/framework/persistentvolumeclaim.go

@@ -66,7 +66,7 @@ func WaitPVCPresentOnClusterFitWith(cluster, namespace, name string, fit func(pv
 			return false
 		}
 		return fit(pvc)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitPVCDisappearOnCluster wait PersistentVolumeClaim disappear on cluster until timeout.
@@ -86,7 +86,7 @@ func WaitPVCDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get PersistentVolumeClaim(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitPVCDisappearOnClusters Wait for the PersistentVolumeClaim to disappear on member clusters until timeout.

test/e2e/framework/pod.go

@@ -64,7 +64,7 @@ func WaitPodPresentOnClusterFitWith(cluster, namespace, name string, fit func(po
 			return false
 		}
 		return fit(pod)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitPodMetricsReady wait podMetrics to be ready.
@@ -90,7 +90,7 @@ func WaitPodMetricsReady(kubeClient kubernetes.Interface, karmadaClient karmada.
 			return false
 		}
 		return true
-	}, metricsCreationDelay, pollInterval).Should(gomega.Equal(true))
+	}, metricsCreationDelay, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitPodPresentOnClustersFitWith wait pod present on cluster sync with fit func.
@@ -119,7 +119,7 @@ func WaitPodDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get pod(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitPodDisappearOnClusters wait pod disappear on member clusters until timeout.

test/e2e/framework/propagationpolicy.go

@@ -94,5 +94,5 @@ func WaitPropagationPolicyFitWith(client karmada.Interface, namespace, name stri
 			return false
 		}
 		return fit(policy)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/rbac.go

@@ -70,7 +70,7 @@ func WaitRolePresentOnClusterFitWith(cluster, namespace, name string, fit func(r
 			return false
 		}
 		return fit(role)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitRoleDisappearOnClusters wait role disappear on member clusters until timeout.
@@ -99,7 +99,7 @@ func WaitRoleDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get role(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // CreateClusterRole create clusterRole.
@@ -142,7 +142,7 @@ func WaitClusterRolePresentOnClusterFitWith(cluster, name string, fit func(clust
 			return false
 		}
 		return fit(clusterRole)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitClusterRoleDisappearOnClusters wait clusterRole disappear on member clusters until timeout.
@@ -171,7 +171,7 @@ func WaitClusterRoleDisappearOnCluster(cluster, name string) {
 
 		klog.Errorf("Failed to get clusterRole(%s) on cluster(%s), err: %v", name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitClusterRoleGetByClientFitWith wait clusterRole get by client fit with func.
@@ -183,7 +183,7 @@ func WaitClusterRoleGetByClientFitWith(client kubernetes.Interface, name string,
 				return false
 			}
 			return fit(clusterRole)
-		}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+		}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 	})
 }
 
@@ -227,7 +227,7 @@ func WaitRoleBindingPresentOnClusterFitWith(cluster, namespace, name string, fit
 			return false
 		}
 		return fit(roleBinding)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitRoleBindingDisappearOnClusters wait roleBinding disappear on member clusters until timeout.
@@ -256,7 +256,7 @@ func WaitRoleBindingDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get roleBinding(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // CreateClusterRoleBinding create clusterRoleBinding.
@@ -299,7 +299,7 @@ func WaitClusterRoleBindingPresentOnClusterFitWith(cluster, name string, fit fun
 			return false
 		}
 		return fit(clusterRoleBinding)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitClusterRoleBindingDisappearOnClusters wait clusterRoleBinding disappear on member clusters until timeout.
@@ -328,7 +328,7 @@ func WaitClusterRoleBindingDisappearOnCluster(cluster, name string) {
 
 		klog.Errorf("Failed to get clusterRoleBinding(%s) on cluster(%s), err: %v", name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // CreateServiceAccount create serviceaccount.
@@ -362,7 +362,7 @@ func WaitServiceAccountPresentOnClusterFitWith(cluster, namespace, name string,
 			return false
 		}
 		return fit(sa)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitServiceAccountPresentOnClustersFitWith wait sa present on cluster sync with fit func.
@@ -391,7 +391,7 @@ func WaitServiceAccountDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get sa(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitServiceAccountDisappearOnClusters wait sa disappear on member clusters until timeout.

test/e2e/framework/resource/operator/karmada.go

@@ -0,0 +1,87 @@
+/*
+Copyright 2025 The Karmada Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package operator
+
+import (
+	"context"
+	"fmt"
+	"time"
+
+	"github.com/onsi/ginkgo/v2"
+	"github.com/onsi/gomega"
+	apierrors "k8s.io/apimachinery/pkg/api/errors"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/klog/v2"
+
+	operatorv1alpha1 "github.com/karmada-io/karmada/operator/pkg/apis/operator/v1alpha1"
+	operator "github.com/karmada-io/karmada/operator/pkg/generated/clientset/versioned"
+	"github.com/karmada-io/karmada/test/e2e/framework"
+)
+
+// WaitKarmadaReady wait karmada instance ready until timeout.
+// Since the karmada-operator updates the `karmada.spec` first and then the `karmada.status`, in order to ensure that the `ready` condition indicates
+// that the `karmada.spec` has been applied correctly, it will check the `lastTransitionTime` of the `ready` condition.
+func WaitKarmadaReady(client operator.Interface, namespace, name string, lastTransitionTime time.Time) {
+	klog.Infof("Waiting for karmada instance %s/%s ready", namespace, name)
+	ginkgo.By(fmt.Sprintf("Waiting for karmada instance %s/%s ready", namespace, name), func() {
+		gomega.Eventually(func(g gomega.Gomega) bool {
+			karmada, err := client.OperatorV1alpha1().Karmadas(namespace).Get(context.TODO(), name, metav1.GetOptions{})
+			g.Expect(err).NotTo(gomega.HaveOccurred())
+			for _, condition := range karmada.Status.Conditions {
+				if condition.Type == "Ready" && condition.Status == "True" && condition.LastTransitionTime.After(lastTransitionTime) {
+					return true
+				}
+			}
+			return false
+		}, framework.PollTimeout, framework.PollInterval).Should(gomega.Equal(true))
+	})
+}
+
+// CreateKarmadaInstance creates a karmada instance.
+func CreateKarmadaInstance(operatorClient operator.Interface, karmada *operatorv1alpha1.Karmada) error {
+	_, err := operatorClient.OperatorV1alpha1().Karmadas(karmada.GetNamespace()).Create(context.TODO(), karmada, metav1.CreateOptions{})
+	if err != nil {
+		if apierrors.IsAlreadyExists(err) {
+			return nil
+		}
+
+		return err
+	}
+	return nil
+}
+
+// UpdateKarmadaInstanceWithSpec updates karmada instance with spec.
+func UpdateKarmadaInstanceWithSpec(client operator.Interface, namespace, name string, karmadaSpec operatorv1alpha1.KarmadaSpec) {
+	ginkgo.By(fmt.Sprintf("Updating Karmada(%s/%s) spec", namespace, name), func() {
+		karmada, err := client.OperatorV1alpha1().Karmadas(namespace).Get(context.TODO(), name, metav1.GetOptions{})
+		gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+
+		karmada.Spec = karmadaSpec
+		_, err = client.OperatorV1alpha1().Karmadas(namespace).Update(context.TODO(), karmada, metav1.UpdateOptions{})
+		gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+	})
+}
+
+// GetLastTransitionTime gets the last transition time of the condition, return time.Now() if not found.
+func GetLastTransitionTime(karmada *operatorv1alpha1.Karmada, conditionType operatorv1alpha1.ConditionType) time.Time {
+	for _, condition := range karmada.Status.Conditions {
+		if condition.Type == string(conditionType) {
+			return condition.LastTransitionTime.Time
+		}
+	}
+	return time.Now()
+}

test/e2e/framework/resourcebinding.go

@@ -38,7 +38,7 @@ func WaitResourceBindingFitWith(client karmada.Interface, namespace, name string
 			return false
 		}
 		return fit(resourceBinding)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // AssertBindingScheduledClusters wait deployment present on member clusters sync with fit func.
@@ -61,7 +61,7 @@ func AssertBindingScheduledClusters(client karmada.Interface, namespace, name st
 				}
 			}
 			return fmt.Errorf("scheduled clusters: %+v, expected possible results: %+v", scheduledClusters, expectedResults)
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -77,6 +77,6 @@ func WaitGracefulEvictionTasksDone(client karmada.Interface, namespace, name str
 				return fmt.Errorf("%d GracefulEvictionTasks is being processing", len(binding.Spec.GracefulEvictionTasks))
 			}
 			return nil
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }

test/e2e/framework/resourcequota.go

@@ -46,7 +46,7 @@ func WaitResourceQuotaPresentOnCluster(cluster, namespace, name string) {
 		_, err := clusterClient.CoreV1().ResourceQuotas(namespace).Get(context.TODO(), name, metav1.GetOptions{})
 		g.Expect(err).NotTo(gomega.HaveOccurred())
 		return true, nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitResourceQuotaDisappearOnClusters wait resourceQuota disappear on clusters until timeout.
@@ -75,5 +75,5 @@ func WaitResourceQuotaDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get resourceQuota(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }

test/e2e/framework/secret.go

@@ -66,7 +66,7 @@ func WaitSecretPresentOnClusterFitWith(cluster, namespace, name string, fit func
 			return false
 		}
 		return fit(secret)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitSecretDisappearOnCluster wait secret disappear on cluster until timeout.
@@ -86,7 +86,7 @@ func WaitSecretDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get secret(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitSecretDisappearOnClusters wait service disappear on member clusters until timeout.

test/e2e/framework/service.go

@@ -59,7 +59,7 @@ func WaitServicePresentOnClusterFitWith(cluster, namespace, name string, fit fun
 			return false
 		}
 		return fit(svc)
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitServicePresentOnClustersFitWith wait service present on cluster sync with fit func.
@@ -88,7 +88,7 @@ func WaitServiceDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get service(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitServiceDisappearOnClusters wait service disappear on member clusters until timeout.

test/e2e/framework/statefulSet.go

@@ -50,6 +50,6 @@ func UpdateStatefulSetReplicas(client kubernetes.Interface, statefulSet *appsv1.
 		gomega.Eventually(func() error {
 			_, err := client.AppsV1().StatefulSets(statefulSet.Namespace).Update(context.TODO(), statefulSet, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }

test/e2e/framework/work.go

@@ -40,5 +40,5 @@ func WaitForWorkToDisappear(client karmada.Interface, namespace, name string) {
 			return fmt.Errorf("failed to get work(%s/%s), err: %w", namespace, name, err)
 		}
 		return nil
-	}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+	}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 }

test/e2e/framework/workload.go

@@ -55,7 +55,7 @@ func UpdateWorkload(client dynamic.Interface, workload *workloadv1alpha1.Workloa
 		gomega.Eventually(func() error {
 			_, err = client.Resource(workloadGVR).Namespace(workload.Namespace).Update(context.TODO(), newUnstructuredObj, metav1.UpdateOptions{}, subresources...)
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -69,7 +69,7 @@ func GetWorkload(client dynamic.Interface, namespace, name string) *workloadv1al
 		gomega.Eventually(func() error {
 			unstructuredObj, err = client.Resource(workloadGVR).Namespace(namespace).Get(context.TODO(), name, metav1.GetOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 
 		err = runtime.DefaultUnstructuredConverter.FromUnstructured(unstructuredObj.UnstructuredContent(), workload)
 		gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
@@ -101,7 +101,7 @@ func WaitWorkloadPresentOnClusterFitWith(cluster, namespace, name string, fit fu
 		err = runtime.DefaultUnstructuredConverter.FromUnstructured(workload.UnstructuredContent(), typedObj)
 		g.Expect(err).ShouldNot(gomega.HaveOccurred())
 		return fit(typedObj), nil
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitWorkloadPresentOnClustersFitWith waits workload present on member clusters sync with fit func.
@@ -130,7 +130,7 @@ func WaitWorkloadDisappearOnCluster(cluster, namespace, name string) {
 
 		klog.Errorf("Failed to get workload(%s/%s) on cluster(%s), err: %v", namespace, name, cluster, err)
 		return false
-	}, pollTimeout, pollInterval).Should(gomega.Equal(true))
+	}, PollTimeout, PollInterval).Should(gomega.Equal(true))
 }
 
 // WaitWorkloadDisappearOnClusters wait workload disappear on member clusters until timeout.

test/e2e/framework/workloadrebalancer.go

@@ -64,7 +64,7 @@ func UpdateWorkloadRebalancer(client karmada.Interface, name string, workloads *
 			}
 			_, err = client.AppsV1alpha1().WorkloadRebalancers().Update(context.TODO(), rebalancer, metav1.UpdateOptions{})
 			return err
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -80,7 +80,7 @@ func WaitRebalancerObservedWorkloads(client karmada.Interface, name string, expe
 				return fmt.Errorf("observedWorkloads: %+v, expectedWorkloads: %+v", rebalancer.Status.ObservedWorkloads, expectedWorkloads)
 			}
 			return nil
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }
 
@@ -96,6 +96,6 @@ func WaitRebalancerDisappear(client karmada.Interface, name string) {
 				return err
 			}
 			return fmt.Errorf("WorkloadRebalancer %s still exist: %+v", name, rebalancer)
-		}, pollTimeout, pollInterval).ShouldNot(gomega.HaveOccurred())
+		}, PollTimeout, PollInterval).ShouldNot(gomega.HaveOccurred())
 	})
 }

test/e2e/suites/operator/priorityclass_test.go

@@ -0,0 +1,82 @@
+/*
+Copyright 2025 The Karmada Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package e2e
+
+import (
+	"context"
+
+	"github.com/onsi/ginkgo/v2"
+	"github.com/onsi/gomega"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/util/rand"
+
+	operatorv1alpha1 "github.com/karmada-io/karmada/operator/pkg/apis/operator/v1alpha1"
+	"github.com/karmada-io/karmada/test/e2e/framework/resource/operator"
+)
+
+var _ = ginkgo.Describe("PriorityClass configuration testing", func() {
+	var karmadaName string
+	var karmadaObject *operatorv1alpha1.Karmada
+	var err error
+
+	ginkgo.Context("PriorityClass configuration testing", func() {
+		ginkgo.BeforeEach(func() {
+			karmadaName = KarmadaInstanceNamePrefix + rand.String(RandomStrLength)
+			InitializeKarmadaInstance(operatorClient, testNamespace, karmadaName)
+		})
+
+		ginkgo.AfterEach(func() {
+			err = operatorClient.OperatorV1alpha1().Karmadas(testNamespace).Delete(context.TODO(), karmadaName, metav1.DeleteOptions{})
+			gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+		})
+
+		ginkgo.It("Custom priorityClass configuration", func() {
+			ginkgo.By("Check if default value is system-node-critical", func() {
+				// take etcd as a representative of StatefulSet.
+				etcd, err := kubeClient.AppsV1().StatefulSets(testNamespace).Get(context.TODO(), karmadaName+"-etcd", metav1.GetOptions{})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+				gomega.Expect(etcd.Spec.Template.Spec.PriorityClassName).Should(gomega.Equal("system-node-critical"))
+
+				// take karmada-apiserver as a representative of Deployment.
+				karmadaApiserver, err := kubeClient.AppsV1().Deployments(testNamespace).Get(context.TODO(), karmadaName+"-apiserver", metav1.GetOptions{})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+				gomega.Expect(karmadaApiserver.Spec.Template.Spec.PriorityClassName).Should(gomega.Equal("system-node-critical"))
+			})
+
+			ginkgo.By("Set priorityClass to system-cluster-critical", func() {
+				karmadaObject, err = operatorClient.OperatorV1alpha1().Karmadas(testNamespace).Get(context.TODO(), karmadaName, metav1.GetOptions{})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+				karmadaObject.Spec.Components.Etcd.Local.PriorityClassName = "system-cluster-critical"
+				karmadaObject.Spec.Components.KarmadaAPIServer.PriorityClassName = "system-cluster-critical"
+				operator.UpdateKarmadaInstanceWithSpec(operatorClient, testNamespace, karmadaName, karmadaObject.Spec)
+				operator.WaitKarmadaReady(operatorClient, testNamespace, karmadaName, operator.GetLastTransitionTime(karmadaObject, operatorv1alpha1.Ready))
+			})
+
+			ginkgo.By("Check if the PriorityClass is applied correctly", func() {
+				// take etcd as a representative of StatefulSet.
+				etcd, err := kubeClient.AppsV1().StatefulSets(testNamespace).Get(context.TODO(), karmadaName+"-etcd", metav1.GetOptions{ResourceVersion: "0"})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+				gomega.Expect(etcd.Spec.Template.Spec.PriorityClassName).Should(gomega.Equal("system-cluster-critical"))
+
+				// take karmada-apiserver as a representative of Deployment.
+				karmadaApiserver, err := kubeClient.AppsV1().Deployments(testNamespace).Get(context.TODO(), karmadaName+"-apiserver", metav1.GetOptions{ResourceVersion: "0"})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+				gomega.Expect(karmadaApiserver.Spec.Template.Spec.PriorityClassName).Should(gomega.Equal("system-cluster-critical"))
+			})
+		})
+	})
+})

test/e2e/suites/operator/status_test.go

@@ -0,0 +1,86 @@
+/*
+Copyright 2025 The Karmada Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package e2e
+
+import (
+	"context"
+
+	"github.com/onsi/ginkgo/v2"
+	"github.com/onsi/gomega"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/util/rand"
+
+	operatorv1alpha1 "github.com/karmada-io/karmada/operator/pkg/apis/operator/v1alpha1"
+	operatorutil "github.com/karmada-io/karmada/operator/pkg/util"
+)
+
+var _ = ginkgo.Describe("Status testing", func() {
+	var karmadaName string
+	var karmadaObject *operatorv1alpha1.Karmada
+	var err error
+
+	ginkgo.Context("Karmada instance status testing", func() {
+		ginkgo.BeforeEach(func() {
+			karmadaName = KarmadaInstanceNamePrefix + rand.String(RandomStrLength)
+			InitializeKarmadaInstance(operatorClient, testNamespace, karmadaName)
+		})
+
+		ginkgo.AfterEach(func() {
+			err = operatorClient.OperatorV1alpha1().Karmadas(testNamespace).Delete(context.TODO(), karmadaName, metav1.DeleteOptions{})
+			gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+		})
+
+		ginkgo.It("Check if the karmada status meets the expectations", func() {
+			ginkgo.By("Get the latest karmada instance", func() {
+				karmadaObject, err = operatorClient.OperatorV1alpha1().Karmadas(testNamespace).Get(context.TODO(), karmadaName, metav1.GetOptions{})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+			})
+
+			ginkgo.By("Check if status.conditions meets the expectations", func() {
+				conditions := karmadaObject.Status.Conditions
+				gomega.Expect(len(conditions)).Should(gomega.BeNumerically(">", 0))
+				// check if the Ready condition is true
+				hasReadyCondition := false
+				for i := range karmadaObject.Status.Conditions {
+					switch karmadaObject.Status.Conditions[i].Type {
+					case string(operatorv1alpha1.Ready):
+						gomega.Expect(karmadaObject.Status.Conditions[i].Status).Should(gomega.Equal(metav1.ConditionTrue))
+						hasReadyCondition = true
+					}
+				}
+				gomega.Expect(hasReadyCondition).Should(gomega.BeTrue())
+			})
+
+			ginkgo.By("Check if the status.SecretRef can ref to the right secret", func() {
+				secretRef := karmadaObject.Status.SecretRef
+				gomega.Expect(secretRef).ShouldNot(gomega.BeNil())
+				gomega.Expect(secretRef.Namespace).Should(gomega.Equal(karmadaObject.GetNamespace()))
+				gomega.Expect(secretRef.Name).Should(gomega.Equal(operatorutil.AdminKarmadaConfigSecretName(karmadaObject.GetName())))
+				_, err := kubeClient.CoreV1().Secrets(secretRef.Namespace).Get(context.TODO(), secretRef.Name, metav1.GetOptions{})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+			})
+
+			ginkgo.By("Check if the status.apiServerService can ref to the right service", func() {
+				apiServerService := karmadaObject.Status.APIServerService
+				gomega.Expect(apiServerService).ShouldNot(gomega.BeNil())
+				gomega.Expect(apiServerService.Name).Should(gomega.Equal(operatorutil.KarmadaAPIServerName(karmadaObject.GetName())))
+				_, err := kubeClient.CoreV1().Services(karmadaObject.GetNamespace()).Get(context.TODO(), apiServerService.Name, metav1.GetOptions{})
+				gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+			})
+		})
+	})
+})

test/e2e/suites/operator/suite_test.go

@@ -0,0 +1,142 @@
+/*
+Copyright 2025 The Karmada Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package e2e
+
+import (
+	"flag"
+	"fmt"
+	"os"
+	"os/exec"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/onsi/ginkgo/v2"
+	"github.com/onsi/gomega"
+	"k8s.io/apimachinery/pkg/util/rand"
+	"k8s.io/client-go/kubernetes"
+	"k8s.io/client-go/rest"
+
+	operator "github.com/karmada-io/karmada/operator/pkg/generated/clientset/versioned"
+	"github.com/karmada-io/karmada/pkg/util"
+	"github.com/karmada-io/karmada/test/e2e/framework"
+	operatorresource "github.com/karmada-io/karmada/test/e2e/framework/resource/operator"
+	"github.com/karmada-io/karmada/test/helper"
+)
+
+const (
+	// RandomStrLength represents the random string length to combine names.
+	RandomStrLength = 5
+	// KarmadaInstanceNamePrefix the prefix of the karmada instance name.
+	KarmadaInstanceNamePrefix = "karmadatest-"
+)
+
+var (
+	// pollInterval defines the interval time for a poll operation.
+	pollInterval time.Duration
+	// pollTimeout defines the time after which the poll operation times out.
+	pollTimeout time.Duration
+)
+
+var (
+	hostContext    string
+	kubeconfig     string
+	karmadactlPath string
+	restConfig     *rest.Config
+	kubeClient     kubernetes.Interface
+	testNamespace  string
+	operatorClient operator.Interface
+)
+
+func init() {
+	// usage ginkgo -- --poll-interval=5s --poll-timeout=5m
+	// eg. ginkgo -v --race --trace --fail-fast -p --randomize-all ./test/e2e/ -- --poll-interval=5s --poll-timeout=5m
+	flag.DurationVar(&pollInterval, "poll-interval", 5*time.Second, "poll-interval defines the interval time for a poll operation")
+	flag.DurationVar(&pollTimeout, "poll-timeout", 300*time.Second, "poll-timeout defines the time which the poll operation times out")
+	flag.StringVar(&hostContext, "host-context", "karmada-host", "Name of the host cluster context in control plane kubeconfig file.")
+}
+
+func TestE2E(t *testing.T) {
+	gomega.RegisterFailHandler(ginkgo.Fail)
+	ginkgo.RunSpecs(t, "E2E Operator Suite")
+}
+
+var _ = ginkgo.SynchronizedBeforeSuite(func() []byte {
+	return nil
+}, func([]byte) {
+	kubeconfig = os.Getenv("KUBECONFIG")
+	gomega.Expect(kubeconfig).ShouldNot(gomega.BeEmpty())
+
+	goPathCmd := exec.Command("go", "env", "GOPATH")
+	goPath, err := goPathCmd.CombinedOutput()
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+
+	formatGoPath := strings.Trim(string(goPath), "\n")
+	karmadactlPath = formatGoPath + "/bin/karmadactl"
+	gomega.Expect(karmadactlPath).ShouldNot(gomega.BeEmpty())
+
+	restConfig, err = framework.LoadRESTClientConfig(kubeconfig, hostContext)
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+
+	kubeClient, err = kubernetes.NewForConfig(restConfig)
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+
+	testNamespace = fmt.Sprintf("operatortest-%s", rand.String(RandomStrLength))
+	err = setupTestNamespace(testNamespace, kubeClient)
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+
+	operatorClient, err = operator.NewForConfig(restConfig)
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+})
+
+var _ = ginkgo.SynchronizedAfterSuite(func() {
+	// cleanup all namespaces we created both in control plane and member clusters.
+	// It will not return error even if there is no such namespace in there that may happen in case setup failed.
+	err := cleanupTestNamespace(testNamespace, kubeClient)
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+}, func() {})
+
+// setupTestNamespace will create a namespace in control plane and all member clusters, most of cases will run against it.
+// The reason why we need a separated namespace is it will make it easier to cleanup resources deployed by the testing.
+func setupTestNamespace(namespace string, kubeClient kubernetes.Interface) error {
+	namespaceObj := helper.NewNamespace(namespace)
+	_, err := util.CreateNamespace(kubeClient, namespaceObj)
+	if err != nil {
+		return err
+	}
+
+	return nil
+}
+
+// cleanupTestNamespace will remove the namespace we set up before for the whole testing.
+func cleanupTestNamespace(namespace string, kubeClient kubernetes.Interface) error {
+	err := util.DeleteNamespace(kubeClient, namespace)
+	if err != nil {
+		return err
+	}
+
+	return nil
+}
+
+// InitializeKarmadaInstance initializes a karmada instance.
+func InitializeKarmadaInstance(client operator.Interface, namespace, name string) {
+	karmada := helper.NewKarmada(namespace, name)
+	now := time.Now()
+	err := operatorresource.CreateKarmadaInstance(client, karmada)
+	gomega.Expect(err).ShouldNot(gomega.HaveOccurred())
+	operatorresource.WaitKarmadaReady(client, namespace, name, now)
+}

test/helper/karmada.go

@@ -0,0 +1,87 @@
+/*
+Copyright 2025 The Karmada Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package helper
+
+import (
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/utils/ptr"
+
+	operatorv1alpha1 "github.com/karmada-io/karmada/operator/pkg/apis/operator/v1alpha1"
+)
+
+// NewKarmada returns a new Karmada instance.
+func NewKarmada(namespace string, name string) *operatorv1alpha1.Karmada {
+	return &operatorv1alpha1.Karmada{
+		ObjectMeta: metav1.ObjectMeta{
+			Namespace: namespace,
+			Name:      name,
+		},
+		Spec: operatorv1alpha1.KarmadaSpec{
+			CRDTarball: &operatorv1alpha1.CRDTarball{
+				HTTPSource: &operatorv1alpha1.HTTPSource{URL: "http://local"},
+			},
+			Components: &operatorv1alpha1.KarmadaComponents{
+				Etcd: &operatorv1alpha1.Etcd{},
+				KarmadaAggregatedAPIServer: &operatorv1alpha1.KarmadaAggregatedAPIServer{
+					CommonSettings: operatorv1alpha1.CommonSettings{
+						Image: operatorv1alpha1.Image{
+							ImageRepository: "docker.io/karmada/karmada-aggregated-apiserver",
+							ImageTag:        "latest",
+						},
+						Replicas: ptr.To[int32](1),
+					},
+				},
+				KarmadaControllerManager: &operatorv1alpha1.KarmadaControllerManager{
+					CommonSettings: operatorv1alpha1.CommonSettings{
+						Image: operatorv1alpha1.Image{
+							ImageRepository: "docker.io/karmada/karmada-controller-manager",
+							ImageTag:        "latest",
+						},
+						Replicas: ptr.To[int32](1),
+					},
+				},
+				KarmadaScheduler: &operatorv1alpha1.KarmadaScheduler{
+					CommonSettings: operatorv1alpha1.CommonSettings{
+						Image: operatorv1alpha1.Image{
+							ImageRepository: "docker.io/karmada/karmada-scheduler",
+							ImageTag:        "latest",
+						},
+						Replicas: ptr.To[int32](1),
+					},
+				},
+				KarmadaWebhook: &operatorv1alpha1.KarmadaWebhook{
+					CommonSettings: operatorv1alpha1.CommonSettings{
+						Image: operatorv1alpha1.Image{
+							ImageRepository: "docker.io/karmada/karmada-webhook",
+							ImageTag:        "latest",
+						},
+						Replicas: ptr.To[int32](1),
+					},
+				},
+				KarmadaMetricsAdapter: &operatorv1alpha1.KarmadaMetricsAdapter{
+					CommonSettings: operatorv1alpha1.CommonSettings{
+						Image: operatorv1alpha1.Image{
+							ImageRepository: "docker.io/karmada/karmada-metrics-adapter",
+							ImageTag:        "latest",
+						},
+						Replicas: ptr.To[int32](1),
+					},
+				},
+			},
+		},
+	}
+}