feat(policies): define ObservabilityK8s policy type with deploy-and-connect and connect-only modes [C9S-121] (#2706)

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

api/portainer.go

@@ -838,6 +838,7 @@ type (
 		NodeCount          int                 `json:"NodeCount" validate:"required"`
 		TotalCPU           int64               `json:"TotalCPU" validate:"required"`
 		TotalMemory        int64               `json:"TotalMemory" validate:"required"`
+		ClusterType        string              `json:"ClusterType,omitempty"`
 		DiagnosticsData    *DiagnosticsData    `json:"DiagnosticsData,omitempty"`
 		PerformanceMetrics *PerformanceMetrics `json:"PerformanceMetrics,omitempty"`
 	}
@@ -2641,6 +2642,7 @@ const (
 	RegistryDocker     PolicyType = "registry-docker"
 	ChangeConfirmation PolicyType = "change-confirmation"
 	CleanupDocker      PolicyType = "cleanup-docker"
+	ObservabilityK8s   PolicyType = "observability-k8s"
 )
 
 type HelmInstallStatus string

pkg/snapshot/kubernetes.go

@@ -66,10 +66,39 @@ func kubernetesSnapshotNodes(snapshot *portainer.KubernetesSnapshot, cli kuberne
 	snapshot.TotalCPU = totalCPUs
 	snapshot.TotalMemory = totalMemory
 	snapshot.NodeCount = len(nodeList.Items)
+	snapshot.ClusterType = clusterTypeFromProviderID(nodeList.Items[0].Spec.ProviderID)
 
 	return nil
 }
 
+const (
+	ClusterTypeGKEAutopilot = "gke-autopilot"
+	ClusterTypeEKSFargate   = "eks-fargate"
+	ClusterTypeAKS          = "aks"
+	ClusterTypeUnknown      = ""
+)
+
+func clusterTypeFromProviderID(providerID string) string {
+	switch {
+	case strings.HasPrefix(providerID, "gce://") && isGKEAutopilotProviderID(providerID):
+		return ClusterTypeGKEAutopilot
+	case strings.HasPrefix(providerID, "aws://") && strings.Contains(strings.ToLower(providerID), "fargate"):
+		return ClusterTypeEKSFargate
+	case strings.HasPrefix(providerID, "azure://"):
+		return ClusterTypeAKS
+	default:
+		return ClusterTypeUnknown
+	}
+}
+
+// isGKEAutopilotProviderID detects GKE Autopilot via the gk3- node-name prefix.
+// ProviderID format: gce://PROJECT/REGION/NODE-NAME
+// Autopilot nodes: gk3-CLUSTER-POOL-SUFFIX; Standard nodes: gke-CLUSTER-POOL-SUFFIX
+func isGKEAutopilotProviderID(providerID string) bool {
+	parts := strings.Split(providerID, "/")
+	return len(parts) > 0 && strings.HasPrefix(parts[len(parts)-1], "gk3-")
+}
+
 // KubernetesSnapshotDiagnostics returns the diagnostics data for the agent
 func KubernetesSnapshotDiagnostics(cli *kubernetes.Clientset, edgeKey string) (*portainer.DiagnosticsData, error) {
 	podID := os.Getenv("HOSTNAME")

pkg/snapshot/kubernetes_test.go

@@ -15,6 +15,29 @@ import (
 	ktesting "k8s.io/client-go/testing"
 )
 
+func TestClusterTypeFromProviderID(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		providerID string
+		expected   string
+	}{
+		{"gce://my-project/us-central1/gk3-my-cluster-pool-abc123", ClusterTypeGKEAutopilot},
+		{"gce://my-project/us-central1/gke-my-cluster-pool-abc123", ClusterTypeUnknown},
+		{"aws:///us-east-1/fargate-12345", ClusterTypeEKSFargate},
+		{"aws:///us-east-1/i-1234567890abcdef0", ClusterTypeUnknown},
+		{"azure:///subscriptions/x/resourceGroups/y/providers/Microsoft.Compute/virtualMachines/NODE", ClusterTypeAKS},
+		{"", ClusterTypeUnknown},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.providerID, func(t *testing.T) {
+			t.Parallel()
+			require.Equal(t, tt.expected, clusterTypeFromProviderID(tt.providerID))
+		})
+	}
+}
+
 func TestKubernetesSnapshotNodes(t *testing.T) {
 	t.Parallel()
 	// Create a fake client
@@ -25,6 +48,9 @@ func TestKubernetesSnapshotNodes(t *testing.T) {
 		ObjectMeta: metav1.ObjectMeta{
 			Name: "test-node-1",
 		},
+		Spec: corev1.NodeSpec{
+			ProviderID: "gce://my-project/us-central1/gk3-my-cluster-pool-abc123",
+		},
 		Status: corev1.NodeStatus{
 			Capacity: corev1.ResourceList{
 				corev1.ResourceCPU:    resource.MustParse("6"),    // 6 CPU cores
@@ -72,10 +98,11 @@ func TestKubernetesSnapshotNodes(t *testing.T) {
 	require.NoError(t, err)
 
 	// Verify the results - these should match what kubernetesSnapshotNodes would produce
-	require.Equal(t, 3, snapshot.NodeCount)                    // 3 nodes
-	require.Equal(t, int64(12), snapshot.TotalCPU)             // 6 + 4 + 2 = 12 CPUs
-	require.Equal(t, int64(25769803776), snapshot.TotalMemory) // 12GB + 8GB + 4GB = 24GB in bytes
-	require.Nil(t, snapshot.PerformanceMetrics)                // Performance metrics are no longer collected server-side
+	require.Equal(t, 3, snapshot.NodeCount)                         // 3 nodes
+	require.Equal(t, int64(12), snapshot.TotalCPU)                  // 6 + 4 + 2 = 12 CPUs
+	require.Equal(t, int64(25769803776), snapshot.TotalMemory)      // 12GB + 8GB + 4GB = 24GB in bytes
+	require.Equal(t, ClusterTypeGKEAutopilot, snapshot.ClusterType) // detected from node1's ProviderID
+	require.Nil(t, snapshot.PerformanceMetrics)                     // Performance metrics are no longer collected server-side
 
 	t.Logf("kubernetesSnapshotNodes test result: Nodes=%d, CPUs=%d, Memory=%d bytes",
 		snapshot.NodeCount, snapshot.TotalCPU, snapshot.TotalMemory)
@@ -94,6 +121,7 @@ func TestKubernetesSnapshotNodesEmptyCluster(t *testing.T) {
 	require.Equal(t, 0, snapshot.NodeCount)
 	require.Equal(t, int64(0), snapshot.TotalCPU)
 	require.Equal(t, int64(0), snapshot.TotalMemory)
+	require.Equal(t, ClusterTypeUnknown, snapshot.ClusterType)
 	require.Nil(t, snapshot.PerformanceMetrics) // Performance metrics should not be set for empty cluster
 
 	t.Log("Empty cluster test passed - no nodes found, early return behavior confirmed")