feat: add dashboard & openwebui to k8s deploy #6
Workflow file for this run
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # Shared configuration for Kubernetes integration tests | ||
|
Check failure on line 1 in .github/workflows/k8s-shared-config.yml
|
||
| # This file contains common environment variables and configurations | ||
| env: | ||
| KIND_VERSION: v0.20.0 | ||
| KUBECTL_VERSION: v1.28.0 | ||
| KUSTOMIZE_VERSION: v5.7.1 | ||
| KUBE_NAMESPACE: vllm-semantic-router-system | ||
| CLUSTER_NAME: semantic-router-cluster | ||
| TEST_IMAGE_TAG: test | ||
| API_TEST_IMAGE_TAG: api-test | ||
| # Common step configurations | ||
| common_steps: | ||
| checkout: | ||
| - name: Checkout code | ||
| uses: actions/checkout@v4 | ||
| install_kustomize: | ||
| - name: Install Kustomize | ||
| run: | | ||
| echo "Installing Kustomize ${{ env.KUSTOMIZE_VERSION }}..." | ||
| # Use the official installation script for better reliability | ||
| curl -s "https://raw.githubusercontent.com/kubernetes-sigs/kustomize/master/hack/install_kustomize.sh" | bash | ||
| sudo mv kustomize /usr/local/bin/ | ||
| kustomize version | ||
| setup_kubectl: | ||
| - name: Setup kubectl | ||
| run: | | ||
| curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl" | ||
| chmod +x kubectl | ||
| sudo mv kubectl /usr/local/bin/ | ||
| cleanup_disk_space: | ||
| - name: Free up disk space before build | ||
| run: | | ||
| echo "=== Initial disk usage ===" | ||
| df -h | ||
| echo "" | ||
| echo "=== Cleaning up system ===" | ||
| # Remove unnecessary packages and caches | ||
| sudo apt-get clean | ||
| sudo apt-get autoremove -y | ||
| sudo rm -rf /var/lib/apt/lists/* | ||
| sudo rm -rf /tmp/* | ||
| sudo rm -rf /var/tmp/* | ||
| # Clean Docker system | ||
| docker system prune -af --volumes | ||
| # Remove large unnecessary files/directories | ||
| sudo rm -rf /usr/share/dotnet | ||
| sudo rm -rf /usr/local/lib/android | ||
| sudo rm -rf /opt/ghc | ||
| sudo rm -rf /opt/hostedtoolcache/CodeQL | ||
| echo "" | ||
| echo "=== Disk usage after cleanup ===" | ||
| df -h | ||
| setup_docker_buildx: | ||
| - name: Set up Docker Buildx | ||
| uses: docker/setup-buildx-action@v3 | ||
| # Common kind cluster configuration | ||
| kind_config: | | ||
| # kind cluster configuration for CI testing | ||
| kind: Cluster | ||
| apiVersion: kind.x-k8s.io/v1alpha4 | ||
| name: semantic-router-cluster | ||
| nodes: | ||
| - role: control-plane | ||
| # Optimized for CI environment with limited resources | ||
| extraPortMappings: | ||
| - containerPort: 30080 | ||
| hostPort: 30080 | ||
| protocol: TCP | ||
| kubeadmConfigPatches: | ||
| - | | ||
| kind: InitConfiguration | ||
| nodeRegistration: | ||
| kubeletExtraArgs: | ||
| # Reduced resource limits for CI | ||
| system-reserved: memory=512Mi,cpu=250m | ||
| kube-reserved: memory=512Mi,cpu=250m | ||
| eviction-hard: memory.available<512Mi,nodefs.available<10% | ||
| - | | ||
| kind: ClusterConfiguration | ||
| apiServer: | ||
| extraArgs: | ||
| max-requests-inflight: "200" | ||
| max-mutating-requests-inflight: "100" | ||
| etcd: | ||
| local: | ||
| extraArgs: | ||
| quota-backend-bytes: "4294967296" # 4GB (reduced from 8GB) | ||
| # Common kustomization overlay template | ||
| kustomization_overlay_template: | | ||
| apiVersion: kustomize.config.k8s.io/v1beta1 | ||
| kind: Kustomization | ||
| # Reference individual files to avoid circular dependency | ||
| resources: | ||
| - ../namespace.yaml | ||
| - ../pv-models.yaml | ||
| - ../deployment.yaml | ||
| - ../service.yaml | ||
| # Generate ConfigMap (same as base) | ||
| configMapGenerator: | ||
| - name: semantic-router-config | ||
| files: | ||
| - ../config.yaml | ||
| - ../tools_db.json | ||
| namespace: vllm-semantic-router-system | ||
| # Use the test image | ||
| images: | ||
| - name: ghcr.io/vllm-project/semantic-router/extproc | ||
| newTag: {IMAGE_TAG} | ||
| # Patch for CI - reduce resources and set imagePullPolicy | ||
| patches: | ||
| - patch: |- | ||
| - op: replace | ||
| path: /spec/template/spec/containers/0/resources/requests/memory | ||
| value: "{MEMORY_REQUEST}" | ||
| - op: replace | ||
| path: /spec/template/spec/containers/0/resources/requests/cpu | ||
| value: "{CPU_REQUEST}" | ||
| - op: replace | ||
| path: /spec/template/spec/containers/0/resources/limits/memory | ||
| value: "{MEMORY_LIMIT}" | ||
| - op: replace | ||
| path: /spec/template/spec/containers/0/resources/limits/cpu | ||
| value: "{CPU_LIMIT}" | ||
| - op: add | ||
| path: /spec/template/spec/containers/0/imagePullPolicy | ||
| value: "IfNotPresent" | ||
| target: | ||
| kind: Deployment | ||
| name: semantic-router | ||
