test: configure pipeline to run scale test and publish data to app insights

.github/workflows/scale-test.yaml

@@ -11,10 +11,9 @@ on:
         description: "AKS Cluster Name"
         required: true
         type: string
-      location:
-        description: "Azure Location"
+      image_tag:
+        description: "Image Tag (if not set, default for this commit will be used)"
         type: string
-        default: ${{ vars.AZURE_LOCATION }}
       num_deployments:
         description: "Number of Traffic Deployments"
         default: 1000
@@ -27,20 +26,8 @@ on:
         description: "Number of Network Policies"
         default: 1000
         type: number
-      num_unique_labels_per_pod: 
-        description: "Number of Unique Labels per Pod"
-        default: 2
-        type: number
-      num_unique_labels_per_deployment:
-        description: "Number of Unique Labels per Deployment"
-        default: 2
-        type: number
-      num_shared_labels_per_pod:
-        description: "Number of Shared Labels per Pod"
-        default: 3
-        type: number
-      delete_labels: 
-        description: "Delete Labels"
+      cleanup: 
+        description: "Clean up environment after test"
         default: true
         type: boolean
 
@@ -54,10 +41,6 @@ on:
         description: "AKS Cluster Name"
         required: true
         type: string
-      location:
-        description: "Azure Location"
-        type: string
-        default: ${{ vars.AZURE_LOCATION }}
       num_deployments:
         description: "Number of Traffic Deployments"
         default: 1000
@@ -70,20 +53,8 @@ on:
         description: "Number of Network Policies"
         default: 1000
         type: number
-      num_unique_labels_per_pod: 
-        description: "Number of Unique Labels per Pod"
-        default: 2
-        type: number
-      num_unique_labels_per_deployment:
-        description: "Number of Unique Labels per Deployment"
-        default: 2
-        type: number
-      num_shared_labels_per_pod:
-        description: "Number of Shared Labels per Pod"
-        default: 3
-        type: number
-      delete_labels: 
-        description: "Delete Labels"
+      cleanup: 
+        description: "Clean up environment after test"
         default: true
         type: boolean
 
@@ -116,17 +87,16 @@ jobs:
       - name: Run Scale Test
         env:
           AZURE_SUBSCRIPTION_ID: ${{ secrets.AZURE_SUBSCRIPTION }}
-          AZURE_LOCATION: ${{ input.location }}
           AZURE_RESOURCE_GROUP: ${{ inputs.resource_group }}
           CLUSTER_NAME: ${{ inputs.cluster_name }}
           NUM_DEPLOYMENTS: ${{ inputs.num_deployments }}
           NUM_REPLICAS: ${{ inputs.num_replicas }}
           NUM_NETPOLS: ${{ inputs.num_netpol }}
-          NUM_UNIQUE_LABELS_PER_POD: ${{ inputs.num_unique_labels_per_pod }}
-          NUM_SHARED_LABELS_PER_POD: ${{ inputs.num_shared_labels_per_pod }}
-          NUM_UNIQUE_LABELS_PER_DEPLOYMENT: ${{ inputs.num_unique_labels_per_deployment }}
-          DELETE_LABELS: ${{ inputs.delete_labels }}
+          CLEANUP: ${{ inputs.cleanup }}
+          IMAGE_REGISTRY: ${{ vars.ACR_NAME }}
+          IMAGE_NAMESPACE: ${{ github.repository }}
+          TAG: ${{ inputs.image_tag }}
         shell: bash
         run: |
           set -euo pipefail
-          go test -v ./test/e2e/. -timeout 300m -tags=scale -count=1  -args -image-tag=$(make version) -image-registry=${{vars.ACR_NAME}} -image-namespace=${{github.repository}}
+          go test -v ./test/e2e/. -timeout 300m -tags=scale -count=1  -args -image-tag=$( [[ $TAG == "" ]] && make version || echo $TAG ) -create-infra=false -delete-infra=false 

go.mod

@@ -333,6 +333,7 @@ require (
 	k8s.io/apiextensions-apiserver v0.30.3
 	k8s.io/cli-runtime v0.30.3
 	k8s.io/kubectl v0.30.3
+	k8s.io/metrics v0.30.3
 	k8s.io/perf-tests/network/benchmarks/netperf v0.0.0-00010101000000-000000000000
 	sigs.k8s.io/controller-runtime v0.18.5
 )

go.sum

@@ -1200,6 +1200,8 @@ k8s.io/kube-openapi v0.0.0-20240903163716-9e1beecbcb38 h1:1dWzkmJrrprYvjGwh9kEUx
 k8s.io/kube-openapi v0.0.0-20240903163716-9e1beecbcb38/go.mod h1:coRQXBK9NxO98XUv3ZD6AK3xzHCxV6+b7lrquKwaKzA=
 k8s.io/kubectl v0.30.3 h1:YIBBvMdTW0xcDpmrOBzcpUVsn+zOgjMYIu7kAq+yqiI=
 k8s.io/kubectl v0.30.3/go.mod h1:IcR0I9RN2+zzTRUa1BzZCm4oM0NLOawE6RzlDvd1Fpo=
+k8s.io/metrics v0.30.3 h1:gKCpte5zykrOmQhZ8qmsxyJslMdiLN+sqbBfIWNpbGM=
+k8s.io/metrics v0.30.3/go.mod h1:W06L2nXRhOwPkFYDJYWdEIS3u6JcJy3ebIPYbndRs6A=
 k8s.io/utils v0.0.0-20240921022957-49e7df575cb6 h1:MDF6h2H/h4tbzmtIKTuctcwZmY0tY9mD9fNT47QO6HI=
 k8s.io/utils v0.0.0-20240921022957-49e7df575cb6/go.mod h1:OLgZIPagt7ERELqWJFomSt595RzquPNLL48iOWgYOg0=
 oras.land/oras-go v1.2.5 h1:XpYuAwAb0DfQsunIyMfeET92emK8km3W4yEzZvUbsTo=

test/e2e/framework/azure/get-fqdn.go

@@ -0,0 +1,27 @@
+package azure
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/Azure/azure-sdk-for-go/sdk/azidentity"
+	armcontainerservice "github.com/Azure/azure-sdk-for-go/sdk/resourcemanager/containerservice/armcontainerservice/v4"
+)
+
+func GetFqdnFn(subscriptionId, resourceGroupName, clusterName string) (string, error) {
+	cred, err := azidentity.NewAzureCLICredential(nil)
+	if err != nil {
+		return "", fmt.Errorf("failed to obtain a credential: %w", err)
+	}
+	ctx := context.Background()
+	clientFactory, err := armcontainerservice.NewClientFactory(subscriptionId, cred, nil)
+	if err != nil {
+		return "", fmt.Errorf("failed to create client: %w", err)
+	}
+	res, err := clientFactory.NewManagedClustersClient().Get(ctx, resourceGroupName, clusterName, nil)
+	if err != nil {
+		return "", fmt.Errorf("failed to finish the get managed cluster client request: %w", err)
+	}
+
+	return *res.Properties.Fqdn, nil
+}

test/e2e/framework/kubernetes/install-retina-helm.go

@@ -19,7 +19,7 @@ import (
 
 const (
 	createTimeout = 20 * time.Minute // windows is slow
-	deleteTimeout = 60 * time.Second
+	deleteTimeout = 5 * time.Minute
 )
 
 var (

test/e2e/framework/kubernetes/uninstall-helm.go

@@ -0,0 +1,45 @@
+package kubernetes
+
+import (
+	"fmt"
+	"log"
+	"os"
+
+	"helm.sh/helm/v3/pkg/action"
+	"helm.sh/helm/v3/pkg/cli"
+)
+
+type UninstallHelmChart struct {
+	Namespace          string
+	ReleaseName        string
+	KubeConfigFilePath string
+}
+
+func (i *UninstallHelmChart) Run() error {
+	settings := cli.New()
+	settings.KubeConfig = i.KubeConfigFilePath
+	actionConfig := new(action.Configuration)
+
+	err := actionConfig.Init(settings.RESTClientGetter(), i.Namespace, os.Getenv("HELM_DRIVER"), log.Printf)
+	if err != nil {
+		return fmt.Errorf("failed to initialize helm action config: %w", err)
+	}
+
+	delclient := action.NewUninstall(actionConfig)
+	delclient.Wait = true
+	delclient.Timeout = deleteTimeout
+	_, err = delclient.Run(i.ReleaseName)
+	if err != nil {
+		return fmt.Errorf("failed to delete existing release %s: %w", i.ReleaseName, err)
+	}
+
+	return nil
+}
+
+func (i *UninstallHelmChart) Prevalidate() error {
+	return nil
+}
+
+func (i *UninstallHelmChart) Stop() error {
+	return nil
+}

test/e2e/framework/scaletest/create-resources.go

@@ -85,7 +85,9 @@ func (c *CreateResources) getResources() []runtime.Object {
 
 	kapingerClusterRoleBinding := kapinger.GetKapingerClusterRoleBinding()
 
-	objs = append(objs, kapingerClusterRole, kapingerClusterRoleBinding)
+	kapingerSA := kapinger.GetKapingerServiceAccount()
+
+	objs = append(objs, kapingerClusterRole, kapingerClusterRoleBinding, kapingerSA)
 	// c.generateKwokNodes()
 	log.Println("Finished generating YAMLs")
 	return objs
@@ -101,6 +103,9 @@ func (c *CreateResources) generateDeployments() []runtime.Object {
 	}
 	template := kapinger.GetKapingerDeployment()
 
+	if template.Labels == nil {
+		template.Labels = make(map[string]string)
+	}
 	template.Labels["is-real"] = "true"
 	template.Spec.Template.Labels["is-real"] = "true"
 	template.Spec.Template.Spec.NodeSelector["scale-test"] = "true"