LiquidCode.Tester/k8s/README.md

Kubernetes Deployment

Prerequisites

• Kubernetes cluster (minikube, kind, or cloud provider)
• kubectl configured
• Docker images built and available

Building Docker Images

# Build Gateway image
docker build -t liquidcode-tester-gateway:latest -f src/LiquidCode.Tester.Gateway/Dockerfile .

# Build C++ Worker image
docker build -t liquidcode-tester-worker-cpp:latest -f src/LiquidCode.Tester.Worker/Dockerfile .

Deploying to Kubernetes

# Create namespace
kubectl apply -f k8s/namespace.yaml

# Apply ConfigMap
kubectl apply -f k8s/configmap.yaml

# Deploy Worker (must be deployed first)
kubectl apply -f k8s/worker-cpp-deployment.yaml

# Deploy Gateway
kubectl apply -f k8s/gateway-deployment.yaml

Checking Status

# Check all resources
kubectl get all -n liquidcode-tester

# Check pods
kubectl get pods -n liquidcode-tester

# Check services
kubectl get services -n liquidcode-tester

# View logs
kubectl logs -n liquidcode-tester -l app=gateway
kubectl logs -n liquidcode-tester -l app=worker-cpp

Access the Gateway

# Get the external IP (for LoadBalancer)
kubectl get service liquidcode-tester-gateway -n liquidcode-tester

# For minikube
minikube service liquidcode-tester-gateway -n liquidcode-tester

# Port forward (alternative)
kubectl port-forward -n liquidcode-tester service/liquidcode-tester-gateway 8080:80

Scaling Workers

# Scale C++ workers
kubectl scale deployment liquidcode-tester-worker-cpp -n liquidcode-tester --replicas=5

Cleanup

# Delete all resources
kubectl delete namespace liquidcode-tester

Production Considerations

1. Image Registry: Push images to a container registry (Docker Hub, GCR, ECR, etc.)
2. Resource Limits: Adjust CPU/Memory limits based on workload
3. Persistent Storage: Add PersistentVolumes for package storage if needed
4. Monitoring: Add Prometheus/Grafana for metrics
5. Logging: Configure centralized logging (ELK, Loki, etc.)
6. Security:
   • Use NetworkPolicies to restrict traffic
   • Enable Pod Security Standards
   • Use secrets for sensitive data
   • Consider using a service mesh (Istio, Linkerd)
7. Autoscaling: Configure HorizontalPodAutoscaler for workers
8. Ingress: Use Ingress controller instead of LoadBalancer for production