fix: stabilize PodCliqueSet hashes against map slice reorder

Makefile

@@ -91,7 +91,7 @@ generate-api-docs: $(CRD_REF_DOCS)
 .PHONY: test-unit
 test-unit:
 	@echo "> Running tests for operator/api"
-	@cd operator/api && go test ./...
+	@make --directory=operator/api test-unit
 	@echo "> Running tests for operator"
 	@make --directory=operator test-unit
 	@echo "> Running tests for operator/client"
@@ -109,12 +109,16 @@ test-unit:
 
 .PHONY: test-cover
 test-cover:
+	@echo "> Running tests with coverage for operator/api"
+	@make --directory=operator/api test-cover
 	@echo "> Running tests with coverage for operator"
 	@make --directory=operator test-cover
 
 # Generates HTML coverage reports for the entire codebase (all modules)
 .PHONY: cover-html
 cover-html:
+	@echo "> Generating HTML coverage report for operator/api"
+	@make --directory=operator/api cover-html
 	@echo "> Generating HTML coverage report for operator"
 	@make --directory=operator cover-html
 

operator/api/Makefile

@@ -35,3 +35,22 @@ generate: $(CONTROLLER_GEN)
 .PHONY: lint
 lint: $(GOLANGCI_LINT)
 	@$(GOLANGCI_LINT) run -c $(REPO_ROOT)/.golangci.yaml ./...
+
+# Make targets for tests
+# -------------------------------------------------------------
+
+# Run all unit tests
+.PHONY: test-unit
+test-unit:
+	@go test ./...
+
+# Run all unit tests with code coverage
+.PHONY: test-cover
+test-cover:
+	@go test ./... -coverprofile=coverage.out
+
+# Generate HTML coverage report
+.PHONY: cover-html
+cover-html: test-cover
+	@go tool cover -html=coverage.out -o coverage.html
+	@echo "Coverage report generated at coverage.html"

operator/api/common/constants/constants.go

@@ -47,6 +47,8 @@ const (
 	// AnnotationDisableManagedResourceProtection is an annotation set by an operator on a PodCliqueSet to explicitly
 	// disable protection of managed resources for a PodCliqueSet.
 	AnnotationDisableManagedResourceProtection = "grove.io/disable-managed-resource-protection"
+	// AnnotationReconcileTrigger is an annotation set on a PodCliqueSet to explicitly trigger a reconcile without changing its spec.
+	AnnotationReconcileTrigger = "grove.io/reconcile-trigger"
 	// AnnotationTopologyName is an annotation set on PodGang to allow KAI scheduler to discover which topology to use.
 	AnnotationTopologyName = "grove.io/topology-name"
 )

operator/e2e/diagnostics/collector.go

@@ -483,6 +483,9 @@ func createDiagnosticsFile(testName, diagDir string) (*os.File, string, error) {
 
 	if diagDir != "" {
 		filename := filepath.Join(diagDir, baseFilename)
+		if err := os.MkdirAll(diagDir, 0755); err != nil {
+			return nil, "", fmt.Errorf("failed to create diagnostics directory %s: %w", diagDir, err)
+		}
 		file, err := os.Create(filename)
 		if err != nil {
 			return nil, "", fmt.Errorf("failed to create diagnostics file in %s: %w", diagDir, err)

operator/e2e/grove/workload/workload.go

@@ -23,6 +23,7 @@ import (
 	"fmt"
 	"time"
 
+	"github.com/ai-dynamo/grove/operator/api/common/constants"
 	grovecorev1alpha1 "github.com/ai-dynamo/grove/operator/api/core/v1alpha1"
 	"github.com/ai-dynamo/grove/operator/e2e/grove/gvk"
 	"github.com/ai-dynamo/grove/operator/e2e/k8s/k8sclient"
@@ -79,7 +80,7 @@ func (wm *WorkloadManager) TriggerPCSReconcile(ctx context.Context, namespace, n
 			Namespace: namespace,
 		},
 	}
-	patch := []byte(fmt.Sprintf(`{"metadata":{"annotations":{"grove.io/reconcile-trigger":%q}}}`, triggerID))
+	patch := []byte(fmt.Sprintf(`{"metadata":{"annotations":{%q:%q}}}`, constants.AnnotationReconcileTrigger, triggerID))
 	if err := wm.cl.Patch(ctx, pcs, client.RawPatch(types.MergePatchType, patch)); err != nil {
 		return fmt.Errorf("trigger PCS reconcile %s/%s: %w", namespace, name, err)
 	}

operator/e2e/testctx/context.go

@@ -190,7 +190,12 @@ func (tc *TestContext) GetLabelSelector() string {
 
 // ListPods lists pods matching the current workload's label selector.
 func (tc *TestContext) ListPods() (*v1.PodList, error) {
-	return tc.newPodManager().List(tc.Ctx, tc.Namespace, tc.GetLabelSelector())
+	return tc.ListPodsWithContext(tc.Ctx)
+}
+
+// ListPodsWithContext lists pods matching the current workload's label selector using ctx.
+func (tc *TestContext) ListPodsWithContext(ctx context.Context) (*v1.PodList, error) {
+	return tc.newPodManager().List(ctx, tc.Namespace, tc.GetLabelSelector())
 }
 
 // WaitForPods waits for the expected pod count to be ready.