test: Run e2e tests in both arm64 and amd64 arch. (#1129)

# Description

This PR allows us to run plugin e2e in both arm64 and amd64 arch.
It also replaces agnhost image with the standard one hosted in
`registry.k8s.io/e2e-test-images`instead of the fork in `acnpublic`, as
it is multi-arch and good practice.

The test sets up agnhost in different args, and check plugin metrics on
retina-agent on both amd64 and arm64.


![image](https://github.com/user-attachments/assets/065ee0f4-a3db-4c07-9c61-1526fc832e1f)


![image](https://github.com/user-attachments/assets/24176641-2e1b-4000-a46f-ef50cc3ace61)



## Related Issue

If this pull request is related to any issue, please mention it here.
Additionally, make sure that the issue is assigned to you before
submitting this pull request.

[#450](#450)

## Checklist

- [x] I have read the [contributing
documentation](https://retina.sh/docs/contributing).
- [x] I signed and signed-off the commits (`git commit -S -s ...`). See
[this
documentation](https://docs.github.com/en/authentication/managing-commit-signature-verification/about-commit-signature-verification)
on signing commits.
- [x] I have correctly attributed the author(s) of the code.
- [x] I have tested the changes locally.
- [x] I have followed the project's style guidelines.
- [x] I have updated the documentation, if necessary.
- [x] I have added tests, if applicable.

## Screenshots (if applicable) or Testing Completed

Please add any relevant screenshots or GIFs to showcase the changes
made.

## Additional Notes

Add any additional notes or context about the pull request here.

---

Please refer to the [CONTRIBUTING.md](../CONTRIBUTING.md) file for more
information on how to contribute to this project.

test/e2e/common/common.go

@@ -26,6 +26,7 @@ const (
 
 var (
 	AzureLocations = []string{"eastus2", "northeurope", "uksouth", "centralindia", "westus2"}
+	Architectures  = []string{"amd64", "arm64"}
 	CreateInfra    = flag.Bool("create-infra", true, "create a Resource group, vNET and AKS cluster for testing")
 	DeleteInfra    = flag.Bool("delete-infra", true, "delete a Resource group, vNET and AKS cluster for testing")
 )

test/e2e/framework/kubernetes/create-agnhost-statefulset.go

@@ -17,15 +17,18 @@ import (
 var ErrLabelMissingFromPod = fmt.Errorf("label missing from pod")
 
 const (
-	AgnhostHTTPPort = 80
-	AgnhostReplicas = 1
+	AgnhostHTTPPort  = 80
+	AgnhostReplicas  = 1
+	AgnhostArchAmd64 = "amd64"
+	AgnhostArchArm64 = "arm64"
 )
 
 type CreateAgnhostStatefulSet struct {
 	AgnhostName        string
 	AgnhostNamespace   string
 	ScheduleOnSameNode bool
 	KubeConfigFilePath string
+	AgnhostArch        string
 }
 
 func (c *CreateAgnhostStatefulSet) Run() error {
@@ -42,14 +45,19 @@ func (c *CreateAgnhostStatefulSet) Run() error {
 	ctx, cancel := context.WithTimeout(context.Background(), defaultTimeoutSeconds*time.Second)
 	defer cancel()
 
-	agnhostStatefulest := c.getAgnhostDeployment()
+	// set default arch to amd64
+	if c.AgnhostArch == "" {
+		c.AgnhostArch = AgnhostArchAmd64
+	}
+
+	agnhostStatefulSet := c.getAgnhostDeployment(c.AgnhostArch)
 
-	err = CreateResource(ctx, agnhostStatefulest, clientset)
+	err = CreateResource(ctx, agnhostStatefulSet, clientset)
 	if err != nil {
 		return fmt.Errorf("error agnhost component: %w", err)
 	}
 
-	selector, exists := agnhostStatefulest.Spec.Selector.MatchLabels["app"]
+	selector, exists := agnhostStatefulSet.Spec.Selector.MatchLabels["app"]
 	if !exists {
 		return fmt.Errorf("missing label \"app=%s\" from agnhost statefulset: %w", c.AgnhostName, ErrLabelMissingFromPod)
 	}
@@ -71,7 +79,7 @@ func (c *CreateAgnhostStatefulSet) Stop() error {
 	return nil
 }
 
-func (c *CreateAgnhostStatefulSet) getAgnhostDeployment() *appsv1.StatefulSet {
+func (c *CreateAgnhostStatefulSet) getAgnhostDeployment(arch string) *appsv1.StatefulSet {
 	reps := int32(AgnhostReplicas)
 
 	var affinity *v1.Affinity
@@ -90,7 +98,6 @@ func (c *CreateAgnhostStatefulSet) getAgnhostDeployment() *appsv1.StatefulSet {
 				},
 			},
 		}
-
 	} else {
 		affinity = &v1.Affinity{
 			PodAntiAffinity: &v1.PodAntiAffinity{
@@ -141,12 +148,13 @@ func (c *CreateAgnhostStatefulSet) getAgnhostDeployment() *appsv1.StatefulSet {
 				Spec: v1.PodSpec{
 					Affinity: affinity,
 					NodeSelector: map[string]string{
-						"kubernetes.io/os": "linux",
+						"kubernetes.io/os":   "linux",
+						"kubernetes.io/arch": arch,
 					},
 					Containers: []v1.Container{
 						{
 							Name:  c.AgnhostName,
-							Image: "acnpublic.azurecr.io/agnhost:2.40",
+							Image: "registry.k8s.io/e2e-test-images/agnhost:2.40",
 							Resources: v1.ResourceRequirements{
 								Requests: v1.ResourceList{
 									"memory": resource.MustParse("20Mi"),

test/e2e/framework/kubernetes/create-network-policy.go

@@ -36,8 +36,8 @@ func (c *CreateDenyAllNetworkPolicy) Run() error {
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 
-	agnhostStatefulest := getNetworkPolicy(c.NetworkPolicyNamespace, c.DenyAllLabelSelector)
-	err = CreateResource(ctx, agnhostStatefulest, clientset)
+	agnhostStatefulSet := getNetworkPolicy(c.NetworkPolicyNamespace, c.DenyAllLabelSelector)
+	err = CreateResource(ctx, agnhostStatefulSet, clientset)
 	if err != nil {
 		return fmt.Errorf("error creating simple deny-all network policy: %w", err)
 	}
@@ -96,8 +96,8 @@ func (d *DeleteDenyAllNetworkPolicy) Run() error {
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 
-	agnhostStatefulest := getNetworkPolicy(d.NetworkPolicyNamespace, d.DenyAllLabelSelector)
-	err = DeleteResource(ctx, agnhostStatefulest, clientset)
+	agnhostStatefulSet := getNetworkPolicy(d.NetworkPolicyNamespace, d.DenyAllLabelSelector)
+	err = DeleteResource(ctx, agnhostStatefulSet, clientset)
 	if err != nil {
 		return fmt.Errorf("error creating simple deny-all network policy: %w", err)
 	}

test/e2e/jobs/jobs.go

@@ -117,12 +117,6 @@ func InstallAndTestRetinaBasicMetrics(kubeConfigFilePath, chartPath string, test
 		TagEnv:             generic.DefaultTagEnv,
 	}, nil)
 
-	job.AddScenario(drop.ValidateDropMetric(testPodNamespace))
-
-	job.AddScenario(tcp.ValidateTCPMetrics(testPodNamespace))
-
-	job.AddScenario(windows.ValidateWindowsBasicMetric())
-
 	dnsScenarios := []struct {
 		name string
 		req  *dns.RequestValidationParams
@@ -164,8 +158,16 @@ func InstallAndTestRetinaBasicMetrics(kubeConfigFilePath, chartPath string, test
 		},
 	}
 
-	for _, scenario := range dnsScenarios {
-		job.AddScenario(dns.ValidateBasicDNSMetrics(scenario.name, scenario.req, scenario.resp, testPodNamespace))
+	for _, arch := range common.Architectures {
+		job.AddScenario(drop.ValidateDropMetric(testPodNamespace, arch))
+		job.AddScenario(tcp.ValidateTCPMetrics(testPodNamespace, arch))
+
+		for _, scenario := range dnsScenarios {
+			name := scenario.name + " - Arch: " + arch
+			job.AddScenario(dns.ValidateBasicDNSMetrics(name, scenario.req, scenario.resp, testPodNamespace, arch))
+		}
+
+		job.AddScenario(windows.ValidateWindowsBasicMetric())
 	}
 
 	job.AddStep(&kubernetes.EnsureStableComponent{
@@ -230,8 +232,11 @@ func UpgradeAndTestRetinaAdvancedMetrics(kubeConfigFilePath, chartPath, valuesFi
 		},
 	}
 
-	for _, scenario := range dnsScenarios {
-		job.AddScenario(dns.ValidateAdvancedDNSMetrics(scenario.name, scenario.req, scenario.resp, kubeConfigFilePath, testPodNamespace))
+	for _, arch := range common.Architectures {
+		for _, scenario := range dnsScenarios {
+			name := scenario.name + " - Arch: " + arch
+			job.AddScenario(dns.ValidateAdvancedDNSMetrics(name, scenario.req, scenario.resp, kubeConfigFilePath, testPodNamespace, arch))
+		}
 	}
 
 	job.AddScenario(latency.ValidateLatencyMetric(testPodNamespace))
@@ -260,55 +265,6 @@ func ValidateHubble(kubeConfigFilePath, chartPath string, testPodNamespace strin
 
 	job.AddScenario(hubble.ValidateHubbleUIService(kubeConfigFilePath))
 
-	job.AddScenario(drop.ValidateDropMetric(testPodNamespace))
-
-	job.AddScenario(tcp.ValidateTCPMetrics(testPodNamespace))
-
-	dnsScenarios := []struct {
-		name string
-		req  *dns.RequestValidationParams
-		resp *dns.ResponseValidationParams
-	}{
-		{
-			name: "Validate basic DNS request and response metrics for a valid domain",
-			req: &dns.RequestValidationParams{
-				NumResponse: "0",
-				Query:       "kubernetes.default.svc.cluster.local.",
-				QueryType:   "A",
-				Command:     "nslookup kubernetes.default",
-				ExpectError: false,
-			},
-			resp: &dns.ResponseValidationParams{
-				NumResponse: "1",
-				Query:       "kubernetes.default.svc.cluster.local.",
-				QueryType:   "A",
-				ReturnCode:  "No Error",
-				Response:    "10.0.0.1",
-			},
-		},
-		{
-			name: "Validate basic DNS request and response metrics for a non-existent domain",
-			req: &dns.RequestValidationParams{
-				NumResponse: "0",
-				Query:       "some.non.existent.domain.",
-				QueryType:   "A",
-				Command:     "nslookup some.non.existent.domain",
-				ExpectError: true,
-			},
-			resp: &dns.ResponseValidationParams{
-				NumResponse: "0",
-				Query:       "some.non.existent.domain.",
-				QueryType:   "A",
-				Response:    dns.EmptyResponse, // hacky way to bypass the framework for now
-				ReturnCode:  "Non-Existent Domain",
-			},
-		},
-	}
-
-	for _, scenario := range dnsScenarios {
-		job.AddScenario(dns.ValidateBasicDNSMetrics(scenario.name, scenario.req, scenario.resp, testPodNamespace))
-	}
-
 	job.AddStep(&kubernetes.EnsureStableComponent{
 		PodNamespace:           common.KubeSystemNamespace,
 		LabelSelector:          "k8s-app=retina",
@@ -365,4 +321,3 @@ func RunPerfTest(kubeConfigFilePath string, chartPath string) *types.Job {
 
 	return job
 }
-

test/e2e/scenarios/dns/scenarios.go

@@ -36,7 +36,7 @@ type ResponseValidationParams struct {
 }
 
 // ValidateBasicDNSMetrics validates basic DNS metrics present in the metrics endpoint
-func ValidateBasicDNSMetrics(scenarioName string, req *RequestValidationParams, resp *ResponseValidationParams, namespace string) *types.Scenario {
+func ValidateBasicDNSMetrics(scenarioName string, req *RequestValidationParams, resp *ResponseValidationParams, namespace, arch string) *types.Scenario {
 	// generate a random ID using rand
 	id := fmt.Sprintf("basic-dns-port-forward-%d", rand.Int()) // nolint:gosec // fine to use math/rand here
 	agnhostName := "agnhost-" + id
@@ -46,6 +46,7 @@ func ValidateBasicDNSMetrics(scenarioName string, req *RequestValidationParams,
 			Step: &kubernetes.CreateAgnhostStatefulSet{
 				AgnhostName:      agnhostName,
 				AgnhostNamespace: namespace,
+				AgnhostArch:      arch,
 			},
 		},
 		// Need this delay to guarantee that the pods will have bpf program attached
@@ -149,7 +150,7 @@ func ValidateBasicDNSMetrics(scenarioName string, req *RequestValidationParams,
 }
 
 // ValidateAdvancedDNSMetrics validates the advanced DNS metrics present in the metrics endpoint
-func ValidateAdvancedDNSMetrics(scenarioName string, req *RequestValidationParams, resp *ResponseValidationParams, kubeConfigFilePath string, namespace string) *types.Scenario {
+func ValidateAdvancedDNSMetrics(scenarioName string, req *RequestValidationParams, resp *ResponseValidationParams, kubeConfigFilePath, namespace, arch string) *types.Scenario {
 	// random ID
 	id := fmt.Sprintf("adv-dns-port-forward-%d", rand.Int()) // nolint:gosec // fine to use math/rand here
 	agnhostName := "agnhost-" + id
@@ -159,6 +160,7 @@ func ValidateAdvancedDNSMetrics(scenarioName string, req *RequestValidationParam
 			Step: &kubernetes.CreateAgnhostStatefulSet{
 				AgnhostName:      agnhostName,
 				AgnhostNamespace: namespace,
+				AgnhostArch:      arch,
 			},
 		},
 		// Need this delay to guarantee that the pods will have bpf program attached

test/e2e/scenarios/drop/scenario.go

@@ -16,8 +16,11 @@ const (
 	IPTableRuleDrop = "IPTABLE_RULE_DROP"
 )
 
-func ValidateDropMetric(namespace string) *types.Scenario {
-	name := "Drop Metrics"
+func ValidateDropMetric(namespace, arch string) *types.Scenario {
+	id := "drop-port-forward-" + arch
+	agnhostName := "agnhost-drop"
+	podName := agnhostName + "-0"
+	name := "Drop Metrics - Arch: " + arch
 	steps := []*types.StepWrapper{
 		{
 			Step: &kubernetes.CreateDenyAllNetworkPolicy{
@@ -28,7 +31,8 @@ func ValidateDropMetric(namespace string) *types.Scenario {
 		{
 			Step: &kubernetes.CreateAgnhostStatefulSet{
 				AgnhostNamespace: namespace,
-				AgnhostName:      "agnhost-drop",
+				AgnhostName:      agnhostName,
+				AgnhostArch:      arch,
 			},
 		},
 		// Need this delay to guarantee that the pods will have bpf program attached
@@ -43,7 +47,7 @@ func ValidateDropMetric(namespace string) *types.Scenario {
 		{
 			Step: &kubernetes.ExecInPod{
 				PodNamespace: namespace,
-				PodName:      "agnhost-drop-0",
+				PodName:      podName,
 				Command:      "curl -s -m 5 bing.com",
 			},
 			Opts: &types.StepOptions{
@@ -59,7 +63,7 @@ func ValidateDropMetric(namespace string) *types.Scenario {
 		{
 			Step: &kubernetes.ExecInPod{
 				PodNamespace: namespace,
-				PodName:      "agnhost-drop-0",
+				PodName:      podName,
 				Command:      "curl -s -m 5 bing.com",
 			},
 			Opts: &types.StepOptions{
@@ -74,24 +78,24 @@ func ValidateDropMetric(namespace string) *types.Scenario {
 				LocalPort:             "10093",
 				RemotePort:            "10093",
 				Endpoint:              "metrics",
-				OptionalLabelAffinity: "app=agnhost-drop", // port forward to a pod on a node that also has this pod with this label, assuming same namespace
+				OptionalLabelAffinity: "app=" + agnhostName, // port forward to a pod on a node that also has this pod with this label, assuming same namespace
 			},
 			Opts: &types.StepOptions{
-				RunInBackgroundWithID: "drop-port-forward",
+				RunInBackgroundWithID: id,
 			},
 		},
 		{
 			Step: &ValidateRetinaDropMetric{
 				PortForwardedRetinaPort: "10093",
-				Source:                  "agnhost-drop",
+				Source:                  agnhostName,
 				Reason:                  IPTableRuleDrop,
 				Direction:               "unknown",
 				Protocol:                UDP,
 			},
 		},
 		{
 			Step: &types.Stop{
-				BackgroundID: "drop-port-forward",
+				BackgroundID: id,
 			},
 		},
 
@@ -108,7 +112,7 @@ func ValidateDropMetric(namespace string) *types.Scenario {
 			Step: &kubernetes.DeleteKubernetesResource{
 				ResourceType:      kubernetes.TypeString(kubernetes.StatefulSet),
 				ResourceNamespace: namespace,
-				ResourceName:      "agnhost-drop",
+				ResourceName:      agnhostName,
 			}, Opts: &types.StepOptions{
 				SkipSavingParametersToJob: true,
 			},

test/e2e/scenarios/tcp/scenario.go

@@ -16,8 +16,11 @@ const (
 	IPTableRuleDrop = "IPTABLE_RULE_DROP"
 )
 
-func ValidateTCPMetrics(namespace string) *types.Scenario {
-	Name := "Flow Metrics"
+func ValidateTCPMetrics(namespace, arch string) *types.Scenario {
+	id := "flow-port-forward-" + arch
+	agnhostName := "agnhost-tcp"
+	podName := agnhostName + "-0"
+	Name := "Flow Metrics - Arch: " + arch
 	Steps := []*types.StepWrapper{
 		{
 			Step: &kubernetes.CreateKapingerDeployment{
@@ -27,13 +30,14 @@ func ValidateTCPMetrics(namespace string) *types.Scenario {
 		},
 		{
 			Step: &kubernetes.CreateAgnhostStatefulSet{
-				AgnhostName:      "agnhost-a",
+				AgnhostName:      agnhostName,
 				AgnhostNamespace: namespace,
+				AgnhostArch:      arch,
 			},
 		},
 		{
 			Step: &kubernetes.ExecInPod{
-				PodName:      "agnhost-a-0",
+				PodName:      podName,
 				PodNamespace: namespace,
 				Command:      "curl -s -m 5 bing.com",
 			}, Opts: &types.StepOptions{
@@ -47,7 +51,7 @@ func ValidateTCPMetrics(namespace string) *types.Scenario {
 		},
 		{
 			Step: &kubernetes.ExecInPod{
-				PodName:      "agnhost-a-0",
+				PodName:      podName,
 				PodNamespace: namespace,
 				Command:      "curl -s -m 5 bing.com",
 			}, Opts: &types.StepOptions{
@@ -61,11 +65,11 @@ func ValidateTCPMetrics(namespace string) *types.Scenario {
 				LocalPort:             "10093",
 				RemotePort:            "10093",
 				Endpoint:              "metrics",
-				OptionalLabelAffinity: "app=agnhost-a", // port forward to a pod on a node that also has this pod with this label, assuming same namespace
+				OptionalLabelAffinity: "app=" + agnhostName, // port forward to a pod on a node that also has this pod with this label, assuming same namespace
 			},
 			Opts: &types.StepOptions{
 				SkipSavingParametersToJob: true,
-				RunInBackgroundWithID:     "drop-flow-forward",
+				RunInBackgroundWithID:     id,
 			},
 		},
 		{
@@ -84,13 +88,13 @@ func ValidateTCPMetrics(namespace string) *types.Scenario {
 		},
 		{
 			Step: &types.Stop{
-				BackgroundID: "drop-flow-forward",
+				BackgroundID: id,
 			},
 		},
 		{
 			Step: &kubernetes.DeleteKubernetesResource{
 				ResourceType:      kubernetes.TypeString(kubernetes.StatefulSet),
-				ResourceName:      "agnhost-a",
+				ResourceName:      agnhostName,
 				ResourceNamespace: namespace,
 			}, Opts: &types.StepOptions{
 				SkipSavingParametersToJob: true,