Fix/ci periodic

.github/actions/kind-cluster-setup/action.yaml

@@ -10,9 +10,11 @@ runs:
   steps:
     - uses: actions/checkout@v4
     - name: Create k8s Kind Cluster
-      uses: helm/kind-action@a1b0e391336a6ee6713a0583f8c6240d70863de3 # @v1.12.0
+      uses: helm/kind-action@v1.12.0
       with:
         cluster_name: ${{ inputs.cluster_name }}
+        wait: 300s
+
     - uses: "google-github-actions/auth@v2"
       with:
         project_id: "sunilarora-fp"

.github/workflows/ci-periodic.yaml

@@ -30,7 +30,7 @@ jobs:
   run-eval:
     if: github.repository == 'GoogleCloudPlatform/kubectl-ai'
     runs-on: ubuntu-latest
-    timeout-minutes: 15
+    timeout-minutes: 12
     # Add "id-token" with the intended permissions.
     permissions:
       contents: "read"
@@ -41,13 +41,36 @@ jobs:
         uses: ./.github/actions/kind-cluster-setup
         with:
           cluster_name: periodic-eval-cluster
+        continue-on-error: false
+        timeout-minutes: 3
       - name: Run an easy eval
         run: |
-          TEST_ARGS="--llm-provider vertexai --models gemini-2.5-pro --concurrency=1 --task-pattern=scale-" ./dev/ci/periodics/run-evals.sh
+          for attempt in 1 2; do
+            echo "=== Evaluation attempt $attempt/2 ==="
+            if timeout 4m bash -c 'TEST_ARGS="--llm-provider vertexai --models gemini-2.5-pro --concurrency=1 --task-pattern=scale-" ./dev/ci/periodics/run-evals.sh'; then
+              echo "Evaluation completed successfully on attempt $attempt"
+              break
+            else
+              echo "Attempt $attempt failed or timed out"
+
+              # Cleanup any hanging processes
+              pkill -f k8s-bench || true
+              pkill -f kubectl-ai || true
+
+              if [ $attempt -eq 2 ]; then
+                echo "❌ Both attempts failed"
+                exit 1
+              else
+                echo "Waiting 10 seconds before retry..."
+                sleep 10
+              fi
+            fi
+          done
       - name: Analyse results
         run: |
           ./dev/ci/periodics/analyze-evals.sh
           cat ${{ github.workspace }}/.build/k8s-bench.md >> ${GITHUB_STEP_SUMMARY}
+
 concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref || github.ref }}
   cancel-in-progress: false