feat(storage): add run_benchmark_tests.sh for automated time-based GC…

.github/workflows/storage-benchmark.yml

@@ -0,0 +1,80 @@
+name: GCS DirectPath Read Benchmark
+
+on:
+  pull_request:
+    paths:
+      - 'packages/google-cloud-storage/**'
+      - '.github/workflows/storage-benchmark.yml'
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  pull-requests: write
+  checks: write
+
+jobs:
+  run-benchmark:
+    name: "GCS Read Microbenchmark"
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.10"
+
+      - name: Authenticate to Google Cloud
+        uses: google-github-actions/auth@v2
+        if: env.HAS_GCP_SECRET == 'true'
+        env:
+          HAS_GCP_SECRET: ${{ secrets.GCP_SA_KEY != '' }}
+        with:
+          credentials_json: ${{ secrets.GCP_SA_KEY }}
+
+      - name: Set up Cloud SDK
+        uses: google-github-actions/setup-gcloud@v2
+        if: env.HAS_GCP_SECRET == 'true'
+        env:
+          HAS_GCP_SECRET: ${{ secrets.GCP_SA_KEY != '' }}
+
+      - name: Package Source Archive
+        run: |
+          tar --exclude='.nox' --exclude='venv_*' --exclude='.pytest_cache' --exclude='__pycache__' --exclude='.git' \
+            -czf /tmp/source.tar.gz -C packages google-cloud-storage
+          ls -lh /tmp/source.tar.gz
+
+      - name: Run Cloud Build Benchmark on High-Bandwidth VM
+        if: env.HAS_GCP_SECRET == 'true'
+        env:
+          HAS_GCP_SECRET: ${{ secrets.GCP_SA_KEY != '' }}
+        run: |
+          BUILD_OUTPUT=$(gcloud builds submit /tmp/source.tar.gz \
+            --project="vaibhavpratap-sdk-test" \
+            --region="us-west4" \
+            --config="packages/google-cloud-storage/cloudbuild/benchmarks-cloudbuild.yaml" \
+            --substitutions=COMMIT_SHA="${{ github.event.pull_request.head.sha || github.sha }}",_PR_NUMBER="${{ github.event.pull_request.number }}",_REPO="${{ github.repository }}" \
+            --format="value(id)")
+          echo "BUILD_ID=$BUILD_OUTPUT" >> $GITHUB_ENV
+          echo "Successfully triggered Cloud Build $BUILD_OUTPUT"
+
+      - name: Fetch Benchmark JSON Result
+        if: env.HAS_GCP_SECRET == 'true'
+        env:
+          HAS_GCP_SECRET: ${{ secrets.GCP_SA_KEY != '' }}
+        run: |
+          mkdir -p /tmp/report
+          gcloud storage cp "gs://vaibhavpratap-sdk-test_cloudbuild/build_results/result_${{ env.BUILD_ID }}.json" /tmp/report/bench_result.json 2>/dev/null || true
+
+      - name: Publish Benchmark Results to PR and Checks Tab
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          python3 packages/google-cloud-storage/cloudbuild/publish_check_run.py \
+            --result-file="/tmp/report/bench_result.json" \
+            --repo="${{ github.repository }}" \
+            --pr-number="${{ github.event.pull_request.number }}" \
+            --commit-sha="${{ github.event.pull_request.head.sha || github.sha }}" \
+            --output-markdown="$GITHUB_STEP_SUMMARY"

packages/google-cloud-storage/cloudbuild/benchmarks-cloudbuild.yaml

@@ -0,0 +1,138 @@
+substitutions:
+  _ZONE: "us-west4-a"
+  _VM_NAME: "gcs-benchmark-runner-us-west4-a"
+  _ULIMIT: "65536"
+  _PROCESSES: "48"
+  _COROS: "1"
+  _FILE_SIZE_MIB: "10240"
+  _CHUNK_SIZE_KIB: "102400"
+  _ROUNDS: "2"
+  _ZONAL_BUCKET: "gcs-read-bench-zb-us-west4-a"
+  _PR_NUMBER: ""
+  _REPO: "shradhakatyal/google-cloud-python"
+
+steps:
+  # Step 0: Generate a temporary SSH key for this build run and register with OS Login
+  - name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
+    id: "generate-ssh-key"
+    entrypoint: "bash"
+    args:
+      - "-c"
+      - |
+        mkdir -p /workspace/.ssh
+        ssh-keygen -t rsa -f /workspace/.ssh/google_compute_engine -N '' -C gcb
+        cat /workspace/.ssh/google_compute_engine.pub > /workspace/gcb_ssh_key.pub
+        gcloud compute os-login ssh-keys add \
+          --key-file=/workspace/.ssh/google_compute_engine.pub \
+          --ttl=1h
+    waitFor: ["-"]
+
+  # Step 1: Package google-cloud-storage directory for direct transfer to VM
+  - name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
+    id: "package-code"
+    entrypoint: "bash"
+    args:
+      - "-c"
+      - |
+        tar --exclude='.nox' --exclude='venv_*' --exclude='.pytest_cache' --exclude='__pycache__' --exclude='.git' \
+          -czf /workspace/google-cloud-storage.tar.gz -C /workspace/packages google-cloud-storage
+    waitFor: ["-"]
+
+  # Step 2: Start the standing high-bandwidth VM
+  - name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
+    id: "start-vm"
+    entrypoint: "bash"
+    args:
+      - "-c"
+      - |
+        echo "Starting standing VM ${_VM_NAME} in zone ${_ZONE}..."
+        gcloud compute instances start "${_VM_NAME}" --zone="${_ZONE}"
+    waitFor: ["-"]
+
+  # Step 3: Run the benchmark directly on the VM via private internal IP SSH, fetch results, and stop the VM
+  - name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
+    id: "run-benchmark-on-vm"
+    entrypoint: "bash"
+    args:
+      - "-c"
+      - |
+        set -e
+        echo "Waiting for VM ${_VM_NAME} to become accessible over internal SSH..."
+        for i in $(seq 1 20); do
+          if gcloud compute ssh "${_VM_NAME}" --zone="${_ZONE}" --internal-ip --ssh-key-file=/workspace/.ssh/google_compute_engine --command="echo VM is ready" 2>/dev/null; then
+            echo "VM internal SSH connection established successfully."
+            break
+          fi
+          echo "Waiting for VM internal SSH availability... (attempt $$i/20)"
+          sleep 10
+        done
+
+        echo "Copying package archive and runner script to VM over internal IP..."
+        gcloud compute scp /workspace/google-cloud-storage.tar.gz \
+          packages/google-cloud-storage/cloudbuild/run_benchmark_tests.sh \
+          packages/google-cloud-storage/cloudbuild/seed_benchmark_objects.py \
+          "${_VM_NAME}":~ --zone="${_ZONE}" --internal-ip --ssh-key-file=/workspace/.ssh/google_compute_engine
+
+        echo "Executing benchmark test suite directly on VM via SSH..."
+        set +e
+        gcloud compute ssh "${_VM_NAME}" --zone="${_ZONE}" --internal-ip --ssh-key-file=/workspace/.ssh/google_compute_engine \
+          --command="tar -xzf google-cloud-storage.tar.gz && cd google-cloud-storage && ulimit -n ${_ULIMIT}; PROCESSES=${_PROCESSES} COROS=${_COROS} FILE_SIZE_MIB=${_FILE_SIZE_MIB} CHUNK_SIZE_KIB=${_CHUNK_SIZE_KIB} ROUNDS=${_ROUNDS} TARGET_BUCKET=${_ZONAL_BUCKET} bash cloudbuild/run_benchmark_tests.sh"
+        TEST_EXIT_CODE=$?
+        set -e
+
+        # Copy JSON report back from VM to Cloud Build workspace
+        mkdir -p /workspace/report
+        echo "Fetching benchmark result JSON from VM..."
+        gcloud compute scp "${_VM_NAME}":~/bench_result.json /workspace/report/bench_result.json \
+          --zone="${_ZONE}" --internal-ip --ssh-key-file=/workspace/.ssh/google_compute_engine 2>/dev/null || true
+
+        # Turn off the standing VM to save quota and cost
+        echo "Stopping VM ${_VM_NAME}..."
+        gcloud compute instances stop "${_VM_NAME}" --zone="${_ZONE}" --quiet || true
+
+        exit $$TEST_EXIT_CODE
+    waitFor:
+      - "start-vm"
+      - "generate-ssh-key"
+      - "package-code"
+
+  # Step 4: Format and publish benchmark performance report
+  - name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
+    id: "publish-benchmark-results"
+    entrypoint: "bash"
+    args:
+      - "-c"
+      - |
+        python3 packages/google-cloud-storage/cloudbuild/publish_check_run.py \
+          --result-file="/workspace/report/bench_result.json" \
+          --repo="${_REPO}" \
+          --commit-sha="${COMMIT_SHA}" \
+          --pr-number="${_PR_NUMBER}" \
+          --build-id="${BUILD_ID}" \
+          --project-id="${PROJECT_ID}" \
+          --region="${LOCATION}" \
+          --vm-name="${_VM_NAME}" \
+          --zonal-bucket="${_ZONAL_BUCKET}"
+    waitFor:
+      - "run-benchmark-on-vm"
+
+  # Step 5: Clean up SSH key from OS Login profile
+  - name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
+    id: "cleanup-ssh-key"
+    entrypoint: "bash"
+    args:
+      - "-c"
+      - |
+        echo "Removing temporary build SSH key from OS Login profile..."
+        gcloud compute os-login ssh-keys remove \
+          --key-file=/workspace/gcb_ssh_key.pub || true
+    waitFor:
+      - "publish-benchmark-results"
+
+timeout: "3600s"
+
+options:
+  logging: CLOUD_LOGGING_ONLY
+  dynamicSubstitutions: true
+  pool:
+    name: "projects/${PROJECT_ID}/locations/us-west4/workerPools/benchmark-worker-pool"