Auto-merge main into odh-release

Author: github-actions[bot]

.github/workflows/maas-bff-tests.yml

@@ -0,0 +1,59 @@
+name: Test / MaaS BFF Tests
+on:
+  push:
+    paths:
+      - "packages/maas/**"
+      - "!LICENSE*"
+      - "!DOCKERFILE*"
+      - "!**.gitignore"
+      - "!**.md"
+
+  pull_request:
+    paths:
+      - "packages/maas/**"
+      - ".github/workflows/**"
+      - "!LICENSE*"
+      - "!DOCKERFILE*"
+      - "!**.gitignore"
+      - "!**.md"
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6.0.1
+
+      - name: Setup Go
+        uses: actions/setup-go@v6
+        with:
+          go-version: "1.24.6"
+
+      - name: Lint
+        working-directory: packages/maas/bff
+        run: make lint
+
+      - name: Build (BFF)
+        working-directory: packages/maas/bff
+        run: make build
+
+      - name: Install envtest assets
+        working-directory: packages/maas/bff
+        run: ./bin/setup-envtest-release-0.17 use 1.29.0 --bin-dir ./bin
+
+      - name: Test (BFF API)
+        working-directory: packages/maas/bff
+        env:
+          ENVTEST_ASSETS_DIR: ${{ github.workspace }}/packages/maas/bff/bin
+        run: go test ./internal/api
+
+      - name: Check if there are uncommitted file changes
+        working-directory: packages/maas
+        run: |
+          clean=$(git status --porcelain)
+          if [[ -z "$clean" ]]; then
+              echo "Empty git status --porcelain: $clean"
+          else
+              echo "Uncommitted file changes detected: $clean"
+              git diff
+              exit 1
+          fi

AGENTS.md

@@ -83,5 +83,6 @@ Before performing certain tasks, read and follow the corresponding specialized r
 | **Contract Tests**    | [docs/agent-rules/contract-tests.md](docs/agent-rules/contract-tests.md)     | When working on contract tests or BFF API validation                         |
 | **Cypress E2E Tests** | [docs/agent-rules/cypress-e2e.md](docs/agent-rules/cypress-e2e.md)           | When creating or modifying E2E tests, Robot Framework migrations             |
 | **Cypress Mock Tests**| [docs/agent-rules/cypress-mock.md](docs/agent-rules/cypress-mock.md)         | When creating or modifying mock/component tests                              |
+| **Unit Tests**        | [docs/agent-rules/unit-tests.md](docs/agent-rules/unit-tests.md)             | When creating or modifying Jest unit tests for utilities, hooks, or components |
 
 **Important**: Always read the relevant rule file before starting the task to ensure you follow the project's conventions and patterns.

OWNERS

@@ -16,6 +16,7 @@ filters:
       - workbenches-approvers
       - gen-ai-approvers
       - nim-serving-approvers
+      - autorag-approvers
 
   # Catchall
   '.*':
@@ -147,3 +148,12 @@ filters:
       - notebooks-reviewers
     labels:
       - 'area/notebooks'
+
+  # AutoRAG
+  'packages/autorag/.*':
+    approvers:
+      - autorag-approvers
+    reviewers:
+      - autorag-reviewers
+    labels:
+      - 'area/autorag'

OWNERS_ALIASES

@@ -64,6 +64,7 @@ aliases:
     - mturley
     - ppadti
     - YuliaKrimerman
+    - Philip-Carneiro
 
   # Model Serving & Model Metrics
   model-serving-metrics-approvers:
@@ -101,6 +102,7 @@ aliases:
     - antowaddle
     - FedeAlonso
     - manosnoam
+    - sridarna
   quality-e2e-testing-reviewers:
     - ConorOM1
     - antowaddle
@@ -148,7 +150,7 @@ aliases:
     - rhdedgar
     - iamemilio
     - s-akhtar-baig
-    
+
   # Notebooks
   notebooks-approvers:
     - andyatmiami
@@ -167,5 +169,29 @@ aliases:
   # NIM / Ecosystem Engineering
   nim-serving-approvers:
     - TomerFi
+    - TheiaSurette
+    - trujillm
+    - mtalvi
+    - xieshenzh
+    - swati-kale
   nim-serving-reviewers:
     - TomerFi
+    - TheiaSurette
+    - trujillm
+    - mtalvi
+    - xieshenzh
+    - swati-kale
+
+  # AutoRAG
+  autorag-approvers:
+    - nickmazzi
+    - chrjones-rh
+  autorag-reviewers:
+    - nickmazzi
+    - chrjones-rh
+    - srtanish1992
+    - daniduong
+    - GAUNSD
+    - jefho-rh
+    - MatthewAThompson
+    - jkyaw

backend/src/__tests__/websocket-proxy.spec.ts

@@ -586,6 +586,36 @@ describe('WebSocket K8s Proxy', () => {
         expect.stringContaining('Unexpected response from K8s API'),
       );
     });
+
+    it('should convert reserved close code 1006 to 1011 when forwarding to client', async () => {
+      // Code 1006 (Abnormal Closure) cannot be sent in a close frame per RFC 6455
+      // The ws module enforces this. We convert it to 1011 (Internal Error).
+      await routeHandler(mockConnection, mockRequest);
+
+      const targetCloseHandler = mockTargetSocket.on.mock.calls.find(
+        (call: any) => call[0] === 'close',
+      )?.[1];
+
+      // K8s API closes with code 1006 (abnormal closure - e.g., network drop)
+      targetCloseHandler(1006, 'Abnormal closure');
+
+      // Client socket should receive 1011 instead of 1006
+      expect(mockSourceSocket.close).toHaveBeenCalledWith(1011, 'Abnormal closure');
+    });
+
+    it('should pass through normal close codes unchanged', async () => {
+      await routeHandler(mockConnection, mockRequest);
+
+      const targetCloseHandler = mockTargetSocket.on.mock.calls.find(
+        (call: any) => call[0] === 'close',
+      )?.[1];
+
+      // K8s API closes with normal code 1000
+      targetCloseHandler(1000, 'Normal closure');
+
+      // Client socket should receive 1000 unchanged
+      expect(mockSourceSocket.close).toHaveBeenCalledWith(1000, 'Normal closure');
+    });
   });
 
   describe('Stale Connection Cleanup', () => {
@@ -602,7 +632,7 @@ describe('WebSocket K8s Proxy', () => {
       // Connection is still active, so no cleanup warning yet
       expect(mockLog.warn).not.toHaveBeenCalledWith(
         expect.anything(),
-        expect.stringContaining('Removing stale connection'),
+        expect.stringContaining('Closing stale websocket connection'),
       );
 
       // Advance time beyond stale threshold
@@ -614,9 +644,78 @@ describe('WebSocket K8s Proxy', () => {
         expect.objectContaining({
           inactivityDuration: expect.any(Number),
         }),
-        expect.stringContaining('Removing stale connection from tracking'),
+        expect.stringContaining('Closing stale websocket connection'),
+      );
+    });
+
+    /**
+     * This test verifies the fix for the zombie connection bug:
+     *
+     * When a connection becomes stale, both websockets are now properly closed
+     * so that the client receives the close event and can reconnect.
+     */
+    it('should close websockets when cleaning up stale connections', async () => {
+      await wssK8sRoutes(mockFastify);
+      routeHandler = (mockFastify.get as jest.Mock).mock.calls[0][2];
+
+      // Create a connection
+      await routeHandler(mockConnection, mockRequest);
+
+      // Verify sockets haven't been closed yet
+      expect(mockSourceSocket.close).not.toHaveBeenCalled();
+      expect(mockTargetSocket.close).not.toHaveBeenCalled();
+
+      // Let connection become stale (advance past threshold + cleanup cycle)
+      jest.advanceTimersByTime(STALE_CONNECTION_MS + 60000);
+
+      // Verify stale cleanup ran
+      expect(mockLog.warn).toHaveBeenCalledWith(
+        expect.objectContaining({
+          inactivityDuration: expect.any(Number),
+        }),
+        expect.stringContaining('Closing stale websocket connection'),
+      );
+
+      // Both websockets should be closed with code 1001 (Going Away)
+      expect(mockSourceSocket.close).toHaveBeenCalledWith(
+        1001,
+        'Connection stale due to inactivity',
+      );
+      expect(mockTargetSocket.close).toHaveBeenCalledWith(
+        1001,
+        'Connection stale due to inactivity',
       );
     });
+
+    /**
+     * This test verifies that heartbeat intervals are properly cleared
+     * when cleaning up stale connections.
+     */
+    it('should clear heartbeat interval when cleaning up stale connections', async () => {
+      await wssK8sRoutes(mockFastify);
+      routeHandler = (mockFastify.get as jest.Mock).mock.calls[0][2];
+
+      // Create a connection
+      await routeHandler(mockConnection, mockRequest);
+
+      // Simulate connection opening to start heartbeat
+      const openHandler = mockTargetSocket.on.mock.calls.find(
+        (call: any) => call[0] === 'open',
+      )?.[1];
+      openHandler();
+
+      // Verify heartbeat is running
+      jest.advanceTimersByTime(HEARTBEAT_INTERVAL_MS);
+      expect(mockTargetSocket.ping).toHaveBeenCalledTimes(1);
+
+      // Advance time beyond stale threshold + cleanup cycle
+      jest.advanceTimersByTime(STALE_CONNECTION_MS + 60000);
+
+      // After cleanup, heartbeat should be stopped - no more pings
+      const pingCountAfterCleanup = mockTargetSocket.ping.mock.calls.length;
+      jest.advanceTimersByTime(HEARTBEAT_INTERVAL_MS * 3);
+      expect(mockTargetSocket.ping.mock.calls.length).toBe(pingCountAfterCleanup);
+    });
   });
 
   describe('Configuration Constants', () => {