update

Signed-off-by: bitliu <[email protected]>

Author: Xunzhuo

.github/workflows/integration-test-ai-gateway.yml renamed to .github/workflows/integration-test-k8s.yml

@@ -1,4 +1,4 @@
-name: Integration Test [AI Gateway]
+name: Integration Test [Kubernetes]
 
 on:
   pull_request:

e2e/go.mod

@@ -3,6 +3,7 @@ module github.com/vllm-project/semantic-router/e2e
 go 1.24
 
 require (
+	k8s.io/api v0.31.0
 	k8s.io/apimachinery v0.31.0
 	k8s.io/client-go v0.31.0
 )
@@ -43,7 +44,6 @@ require (
 	gopkg.in/inf.v0 v0.9.1 // indirect
 	gopkg.in/yaml.v2 v2.4.0 // indirect
 	gopkg.in/yaml.v3 v3.0.1 // indirect
-	k8s.io/api v0.31.0 // indirect
 	k8s.io/klog/v2 v2.130.1 // indirect
 	k8s.io/kube-openapi v0.0.0-20240228011516-70dd3763d340 // indirect
 	k8s.io/utils v0.0.0-20240711033017-18e509b52bc8 // indirect

e2e/pkg/framework/runner.go

@@ -263,11 +263,20 @@ func (r *Runner) runSingleTest(ctx context.Context, kubeClient *kubernetes.Clien
 
 	start := time.Now()
 
+	// Get service configuration from profile
+	svcConfig := r.profile.GetServiceConfig()
+
 	opts := testcases.TestCaseOptions{
 		Verbose:    r.opts.Verbose,
 		Namespace:  "default",
 		Timeout:    "5m",
 		RestConfig: r.restConfig,
+		ServiceConfig: testcases.ServiceConfig{
+			LabelSelector: svcConfig.LabelSelector,
+			Namespace:     svcConfig.Namespace,
+			Name:          svcConfig.Name,
+			PortMapping:   svcConfig.PortMapping,
+		},
 	}
 
 	err := tc.Fn(ctx, kubeClient, opts)

e2e/pkg/framework/types.go

@@ -22,6 +22,17 @@ type Profile interface {
 
 	// GetTestCases returns the list of test cases to run for this profile
 	GetTestCases() []string
+
+	// GetServiceConfig returns the service configuration for accessing the deployed service
+	GetServiceConfig() ServiceConfig
+}
+
+// ServiceConfig contains configuration for accessing a service
+type ServiceConfig struct {
+	LabelSelector string // e.g., "gateway.envoyproxy.io/owning-gateway-namespace=default,..."
+	Namespace     string
+	Name          string // optional, if empty will use LabelSelector to find service
+	PortMapping   string // e.g., "8080:80"
 }
 
 // SetupOptions contains options for profile setup

e2e/pkg/helpers/kubernetes.go

@@ -122,11 +122,12 @@ func VerifyServicePodsRunning(ctx context.Context, client *kubernetes.Clientset,
 // The ports parameter should be in format "localPort:servicePort" (e.g., "8080:80")
 // Note: Kubernetes API doesn't support port-forward directly to services, only to pods.
 // This function mimics kubectl's behavior by finding a pod behind the service.
-func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restConfig *rest.Config, namespace, service, ports string, verbose bool) error {
+// Returns a stop function that should be called to clean up the port forwarding.
+func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restConfig *rest.Config, namespace, service, ports string, verbose bool) (func(), error) {
 	// Parse ports (e.g., "8080:80" -> local=8080, service=80)
 	portParts := strings.Split(ports, ":")
 	if len(portParts) != 2 {
-		return fmt.Errorf("invalid port format: %s (expected format: localPort:servicePort)", ports)
+		return nil, fmt.Errorf("invalid port format: %s (expected format: localPort:servicePort)", ports)
 	}
 	localPort := portParts[0]
 	servicePort := portParts[1]
@@ -138,7 +139,7 @@ func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restCon
 	// Get the service to find its selector
 	svc, err := client.CoreV1().Services(namespace).Get(ctx, service, metav1.GetOptions{})
 	if err != nil {
-		return fmt.Errorf("failed to get service: %w", err)
+		return nil, fmt.Errorf("failed to get service: %w", err)
 	}
 
 	// Build label selector from service selector
@@ -153,11 +154,11 @@ func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restCon
 		LabelSelector: labelSelector,
 	})
 	if err != nil {
-		return fmt.Errorf("failed to list pods for service: %w", err)
+		return nil, fmt.Errorf("failed to list pods for service: %w", err)
 	}
 
 	if len(pods.Items) == 0 {
-		return fmt.Errorf("no pods found for service %s/%s", namespace, service)
+		return nil, fmt.Errorf("no pods found for service %s/%s", namespace, service)
 	}
 
 	// Use the first running pod
@@ -171,7 +172,7 @@ func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restCon
 	}
 
 	if targetPod == nil {
-		return fmt.Errorf("no running pods found for service %s/%s", namespace, service)
+		return nil, fmt.Errorf("no running pods found for service %s/%s", namespace, service)
 	}
 
 	if verbose {
@@ -197,7 +198,7 @@ func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restCon
 	// Create SPDY transport
 	transport, upgrader, err := spdy.RoundTripperFor(restConfig)
 	if err != nil {
-		return fmt.Errorf("failed to create SPDY transport: %w", err)
+		return nil, fmt.Errorf("failed to create SPDY transport: %w", err)
 	}
 
 	// Build the URL for port forwarding to pod
@@ -224,7 +225,7 @@ func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restCon
 	// Create port forwarder (forward localPort to containerPort on the pod)
 	forwarder, err := portforward.New(dialer, []string{fmt.Sprintf("%s:%s", localPort, containerPort)}, stopChan, readyChan, out, errOut)
 	if err != nil {
-		return fmt.Errorf("failed to create port forwarder: %w", err)
+		return nil, fmt.Errorf("failed to create port forwarder: %w", err)
 	}
 
 	// Start port forwarding in background
@@ -242,12 +243,19 @@ func StartPortForward(ctx context.Context, client *kubernetes.Clientset, restCon
 		if verbose {
 			fmt.Printf("[Helper] Port forwarding is ready\n")
 		}
-		return nil
+		// Return stop function
+		stopFunc := func() {
+			if verbose {
+				fmt.Printf("[Helper] Stopping port forwarding to %s/%s\n", namespace, service)
+			}
+			close(stopChan)
+		}
+		return stopFunc, nil
 	case <-time.After(30 * time.Second):
 		close(stopChan)
-		return fmt.Errorf("timeout waiting for port forward to be ready")
+		return nil, fmt.Errorf("timeout waiting for port forward to be ready")
 	case <-ctx.Done():
 		close(stopChan)
-		return ctx.Err()
+		return nil, ctx.Err()
 	}
 }

e2e/pkg/testcases/registry.go

@@ -40,6 +40,26 @@ type TestCaseOptions struct {
 
 	// RestConfig is the Kubernetes REST config
 	RestConfig *rest.Config
+
+	// ServiceConfig contains service-specific configuration
+	ServiceConfig ServiceConfig
+}
+
+// ServiceConfig contains configuration for accessing a service
+type ServiceConfig struct {
+	// LabelSelector is the label selector to find the service
+	// Example: "gateway.envoyproxy.io/owning-gateway-namespace=default,gateway.envoyproxy.io/owning-gateway-name=semantic-router"
+	LabelSelector string
+
+	// Namespace is the namespace where the service is located
+	Namespace string
+
+	// Name is the service name (optional, if not using label selector)
+	Name string
+
+	// PortMapping is the port mapping for port-forwarding
+	// Format: "localPort:servicePort" (e.g., "8080:80")
+	PortMapping string
 }
 
 var (

e2e/profiles/ai-gateway/basic_health_check.go

@@ -13,6 +13,9 @@ import (
 	"github.com/vllm-project/semantic-router/e2e/pkg/framework"
 	"github.com/vllm-project/semantic-router/e2e/pkg/helm"
 	"github.com/vllm-project/semantic-router/e2e/pkg/helpers"
+
+	// Import testcases package to register all test cases via their init() functions
+	_ "github.com/vllm-project/semantic-router/e2e/testcases"
 )
 
 // Profile implements the AI Gateway test profile
@@ -104,7 +107,6 @@ func (p *Profile) Teardown(ctx context.Context, opts *framework.TeardownOptions)
 // GetTestCases returns the list of test cases for this profile
 func (p *Profile) GetTestCases() []string {
 	return []string{
-		"basic-health-check",
 		"chat-completions-request",
 		"domain-classify",
 		"cache",
@@ -113,6 +115,15 @@ func (p *Profile) GetTestCases() []string {
 	}
 }
 
+// GetServiceConfig returns the service configuration for accessing the deployed service
+func (p *Profile) GetServiceConfig() framework.ServiceConfig {
+	return framework.ServiceConfig{
+		LabelSelector: "gateway.envoyproxy.io/owning-gateway-namespace=default,gateway.envoyproxy.io/owning-gateway-name=semantic-router",
+		Namespace:     "envoy-gateway-system",
+		PortMapping:   "8080:80",
+	}
+}
+
 func (p *Profile) deploySemanticRouter(ctx context.Context, deployer *helm.Deployer, opts *framework.SetupOptions) error {
 	// Use local Helm chart instead of remote OCI registry
 	chartPath := "deploy/helm/semantic-router"
@@ -263,6 +274,26 @@ func (p *Profile) verifyEnvironment(ctx context.Context, opts *framework.SetupOp
 		}
 	}
 
+	// Check all deployments are healthy
+	p.log("Verifying all deployments are healthy...")
+
+	// Check semantic-router deployment
+	if err := helpers.CheckDeployment(ctx, client, "vllm-semantic-router-system", "semantic-router", p.verbose); err != nil {
+		return fmt.Errorf("semantic-router deployment not healthy: %w", err)
+	}
+
+	// Check envoy-gateway deployment
+	if err := helpers.CheckDeployment(ctx, client, "envoy-gateway-system", "envoy-gateway", p.verbose); err != nil {
+		return fmt.Errorf("envoy-gateway deployment not healthy: %w", err)
+	}
+
+	// Check ai-gateway-controller deployment
+	if err := helpers.CheckDeployment(ctx, client, "envoy-ai-gateway-system", "ai-gateway-controller", p.verbose); err != nil {
+		return fmt.Errorf("ai-gateway-controller deployment not healthy: %w", err)
+	}
+
+	p.log("All deployments are healthy")
+
 	return nil
 }