Fabric-Analytics-MCP

# Microsoft Fabric Analytics MCP Server - Docker Installation ## Quick Start with Docker ### Using Pre-built Image (Coming Soon) ```bash # Pull and run the official image docker run -d \ --name fabric-mcp \ -p 3000:3000 \ -e FABRIC_AUTH_METHOD=bearer_token \ -e FABRIC_CLIENT_ID=your-client-id \ -e FABRIC_CLIENT_SECRET=your-secret \ -e FABRIC_TENANT_ID=your-tenant-id \ santhoshravindran7/fabric-analytics-mcp:latest ``` ### Building from Source ```bash # Clone repository git clone https://github.com/santhoshravindran7/Fabric-Analytics-MCP.git cd Fabric-Analytics-MCP # Build Docker image docker build -t fabric-analytics-mcp . # Run container docker run -d \ --name fabric-mcp \ -p 3000:3000 \ --env-file .env \ fabric-analytics-mcp ``` ### Docker Compose ```yaml version: '3.8' services: fabric-mcp: build: . ports: - "3000:3000" environment: - FABRIC_AUTH_METHOD=bearer_token - FABRIC_CLIENT_ID=${FABRIC_CLIENT_ID} - FABRIC_CLIENT_SECRET=${FABRIC_CLIENT_SECRET} - FABRIC_TENANT_ID=${FABRIC_TENANT_ID} volumes: - ./config:/app/config restart: unless-stopped healthcheck: test: ["CMD", "curl", "-f", "http://localhost:3000/health"] interval: 30s timeout: 10s retries: 3 ``` ### Environment Configuration Create a `.env` file: ```env FABRIC_AUTH_METHOD=bearer_token FABRIC_CLIENT_ID=your-client-id FABRIC_CLIENT_SECRET=your-client-secret FABRIC_TENANT_ID=your-tenant-id FABRIC_DEFAULT_WORKSPACE_ID=your-workspace-id ``` ### Health Check ```bash # Check container health docker exec fabric-mcp curl -f http://localhost:3000/health # View logs docker logs fabric-mcp # Access container shell docker exec -it fabric-mcp bash ``` ## Kubernetes Deployment ### Helm Chart (Recommended) ```bash # Add Helm repository (coming soon) helm repo add fabric-mcp https://santhoshravindran7.github.io/fabric-analytics-mcp-helm # Install with custom values helm install fabric-mcp fabric-mcp/fabric-analytics-mcp \ --set auth.method=service_principal \ --set auth.clientId=your-client-id \ --set auth.clientSecret=your-secret \ --set auth.tenantId=your-tenant-id ``` ### Manual Kubernetes Deployment ```bash # Apply all Kubernetes manifests kubectl apply -f k8s/ # Check deployment status kubectl get pods -l app=fabric-analytics-mcp # Port forward for local access kubectl port-forward service/fabric-analytics-mcp 3000:3000 ``` ### Configuration via ConfigMap ```yaml apiVersion: v1 kind: ConfigMap metadata: name: fabric-mcp-config data: FABRIC_AUTH_METHOD: "service_principal" FABRIC_DEFAULT_WORKSPACE_ID: "your-workspace-id" ``` ## Production Considerations ### Security - Use Kubernetes secrets for sensitive data - Enable TLS/SSL for external access - Configure network policies - Use service accounts with minimal permissions ### Scaling - Configure horizontal pod autoscaling (HPA) - Set appropriate resource limits and requests - Consider using node affinity for performance ### Monitoring - Prometheus metrics enabled by default - Grafana dashboards included - Health check endpoints available - Structured logging with correlation IDs ### Backup & Recovery - Configuration stored in ConfigMaps - No persistent storage required for stateless operation - Consider backing up logs for audit purposes