[CI/Build] : add GitHub Actions workflows for router (#74) (#94)

* feat(ci): add GitHub Actions workflows for router (#74)
- Add router-e2e-test workflow for PR validation with performance tests
- Add router-docker-release workflow for automated image builds
- Enhance path handling in minikube and router scripts
- Update workflow configurations for better reliability
- Fix Python dependencies and test requirements
- Improve test environment setup and execution

Signed-off-by: Sozhan Natarajan <[email protected]>

.github/workflows/router-docker-release.yml

@@ -0,0 +1,47 @@
+name: Release vLLM Router Docker Image
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - 'src/vllm_router/**'
+      - 'docker/**'
+      - 'setup.py'
+
+jobs:
+  release:
+    permissions:
+      contents: write
+      packages: write
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract version
+        id: version
+        run: echo "version=$(python setup.py --version)" >> "$GITHUB_OUTPUT"
+      - name: Build and push image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: docker/Dockerfile
+          push: true
+          tags: |
+            ghcr.io/${{ github.repository }}/router:latest
+            ghcr.io/${{ github.repository }}/router:${{ steps.version.outputs.version }}
+          cache-from: type=registry,ref=ghcr.io/${{ github.repository }}/router:buildcache
+          cache-to: type=registry,ref=ghcr.io/${{ github.repository }}/router:buildcache,mode=max

.github/workflows/router-e2e-test.yml

@@ -0,0 +1,106 @@
+name: Router E2E tests
+
+on:
+  pull_request:
+    types: [ready_for_review]
+    paths:
+      - 'src/vllm_router/**'
+      - 'docker/**'
+      - 'setup.py'
+  workflow_dispatch:
+
+jobs:
+  e2e-test:
+    runs-on: ubuntu-latest
+    if: github.event.pull_request.draft == false
+    permissions:
+      contents: read
+      packages: write
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Setup Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.10.13"
+          cache: pip
+
+      - name: Login to Github Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -r src/tests/requirements.txt
+          pip install -r requirements-test.txt
+          pip install -e .
+
+      - name: Make scripts executable
+        run: |
+          chmod +x ./utils/install-minikube-cluster.sh
+          chmod +x ./utils/install-kubectl.sh
+          chmod +x ./utils/install-helm.sh
+          chmod +x src/vllm_router/perf-test.sh
+          chmod +x src/tests/perftest/*.sh
+
+      - name: Setup Test environment
+        run: |
+          ./utils/install-minikube-cluster.sh
+
+      - name: Build and Load test Image
+        run: |
+          REPO=$(echo "${{ github.repository }}" | tr '[:upper:]' '[:lower:]')
+          docker build -t "ghcr.io/$REPO/router:test" -f docker/Dockerfile .
+          minikube image load "ghcr.io/$REPO/router:test"
+
+      - name: Start Mock OpenAI servers
+        working-directory: src/tests/perftest
+        run: |
+          bash run-multi-server.sh 4 500
+          sleep 10
+
+      - name: Start Router for Testing
+        run: |
+          bash src/vllm_router/perf-test.sh 8000 &
+          sleep 5
+
+      - name: Run Performance tests
+        working-directory: src/tests/perftest
+        env:
+          PYTHONPATH: ${{ github.workspace }}
+        run: |
+          echo "PYTHONPATH=$PYTHONPATH"
+          mkdir -p logs
+          if [ ! -f "request_generator.py" ]; then
+            echo "Error: request_generator.py not found!"
+            exit 1
+          fi
+          # Run with Python debug option
+          python3 -v request_generator.py --qps 10 --num-workers 32 --duration 300 2>&1 | tee logs/request_generator.log
+
+      - name: Run E2E Tests
+        run: |
+          pytest src/tests/test_*.py
+
+      - name: Cleanup Test Environment
+        if: always()
+        working-directory: src/tests/perftest
+        run: |
+          bash clean-up.sh
+
+      - name: Upload Test logs
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: test-logs-pr-${{ github.event.pull_request.number || 'manual' }}
+          path: |
+            ~/.kube/config
+            /tmp/minikube.log
+            src/tests/perftest/logs

src/tests/requirements.txt

@@ -1,3 +1,4 @@
 fastapi
 httpx
 uvicorn
+vllm

src/vllm_router/perf-test.sh

@@ -4,7 +4,11 @@ if [[ $# -ne 1 ]]; then
     exit 1
 fi
 
-python3 router.py --port "$1" \
+# Get the directory where the script is located
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+# Run router.py from the correct directory
+python3 "$SCRIPT_DIR/router.py" --port "$1" \
     --service-discovery static \
     --static-backends "http://localhost:9004,http://localhost:9001,http://localhost:9002,http://localhost:9003" \
     --static-models "fake_model_name,fake_model_name,fake_model_name,fake_model_name" \

utils/install-minikube-cluster.sh

@@ -5,9 +5,12 @@ minikube_exists() {
   command -v minikube >/dev/null 2>&1
 }
 
+# Get script directory for relative paths
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
 # Install kubectl and helm
-bash ./install-kubectl.sh
-bash ./install-helm.sh
+bash "$SCRIPT_DIR/install-kubectl.sh"
+bash "$SCRIPT_DIR/install-helm.sh"
 
 # Install minikube
 if minikube_exists; then
@@ -17,20 +20,32 @@ else
   sudo install minikube-linux-amd64 /usr/local/bin/minikube && rm minikube-linux-amd64
 fi
 
-echo "net.core.bpf_jit_harden=0" | sudo tee -a /etc/sysctl.conf
-sudo sysctl -p
-
-# Install nvidia-container-toolkit
-sudo nvidia-ctk runtime configure --runtime=docker && sudo systemctl restart docker
+# Configure BPF if available
+if [ -f /proc/sys/net/core/bpf_jit_harden ]; then
+    echo "net.core.bpf_jit_harden=0" | sudo tee -a /etc/sysctl.conf
+    sudo sysctl -p
+else
+    echo "BPF JIT hardening configuration not available, skipping..."
+fi
 
-# Start cluster
-sudo minikube start --driver docker --container-runtime docker --gpus all --force --addons=nvidia-device-plugin
+# Check if NVIDIA GPU is available
+if command -v nvidia-smi &> /dev/null; then
+    # Install nvidia-container-toolkit
+    sudo nvidia-ctk runtime configure --runtime=docker && sudo systemctl restart docker
 
-# Install gpu-operator
-sudo helm repo add nvidia https://helm.ngc.nvidia.com/nvidia \
-    && sudo helm repo update
+    # Start cluster with GPU support
+    minikube start --driver docker --container-runtime docker --gpus all --force --addons=nvidia-device-plugin
 
-sudo helm install --wait --generate-name \
-    -n gpu-operator --create-namespace \
-    nvidia/gpu-operator \
-    --version=v24.9.1
+    # Install gpu-operator
+    sudo helm repo add nvidia https://helm.ngc.nvidia.com/nvidia && sudo helm repo update
+    sudo helm install --wait --generate-name \
+        -n gpu-operator --create-namespace \
+        nvidia/gpu-operator \
+        --version=v24.9.1
+else
+    echo "No NVIDIA GPU detected, starting minikube without GPU support..."
+    # Fix permission issues
+    sudo sysctl fs.protected_regular=0
+    # Start cluster without GPU
+    minikube start --driver docker --force
+fi