DevOps AI Toolkit

# Docker Setup Guide **Complete Docker deployment with MCP server + Qdrant vector database - all features included out-of-the-box.** ## When to Use This Method ✅ **Perfect for:** - Individual developers running MCP locally - Full-featured setup with all capabilities working immediately - Local development and testing environments - Users who want everything working without external dependencies ❌ **Consider alternatives for:** - Company-wide or centralized deployments (use Docker with external orchestration) - Environments without Docker → See [other setup methods](../mcp-setup.md#setup-methods) for alternatives ## What You Get - **DevOps AI Toolkit MCP Server** - All Kubernetes deployment and testing features - **Qdrant Vector Database** - Automatic setup for capability and pattern management - **Complete Integration** - All features work immediately - **Isolated Local Environment** - No conflicts with local Node.js or other services - **Local Development Focus** - Perfect for individual developer workflows ## Prerequisites - Docker and Docker Compose installed - Kubernetes cluster access (kubectl configured) - AI model API key (default: Anthropic). See [AI Model Configuration](../mcp-setup.md#ai-model-configuration) for available model options. - OpenAI API key (for enhanced semantic search) ## Quick Start (2 Minutes) ### 1. Download Docker Compose Configuration ```bash # Download Docker Compose configuration curl -o docker-compose-dot-ai.yaml https://raw.githubusercontent.com/vfarcic/dot-ai/main/docker-compose-dot-ai.yaml ``` ### 2. Set Environment Variables **All environment variables from the [Configuration Overview](../mcp-setup.md#configuration-overview) can be used with Docker setup.** Variables must be added to the `docker-compose-dot-ai.yaml` file's `environment:` section to be passed to the container. **Quick start example:** ```bash # Required: Anthropic API key export ANTHROPIC_API_KEY="sk-ant-api03-..." # Required: OpenAI API key for enhanced semantic search export OPENAI_API_KEY="sk-proj-..." # Optional: Custom Docker images and ports (defaults shown) export DOT_AI_IMAGE="ghcr.io/vfarcic/dot-ai:0.68.0" # See available versions: https://github.com/vfarcic/dot-ai/pkgs/container/dot-ai export QDRANT_IMAGE="qdrant/qdrant:v1.7.4" export QDRANT_PORT="6333" # Optional: Custom kubeconfig path (defaults to ~/.kube/config) export KUBECONFIG="/path/to/your/kubeconfig.yaml" # Optional: Custom endpoints (OpenRouter, self-hosted) # See: https://github.com/vfarcic/dot-ai/blob/main/docs/mcp-setup.md#custom-endpoint-configuration export CUSTOM_LLM_API_KEY="sk-or-v1-..." export CUSTOM_LLM_BASE_URL="https://openrouter.ai/api/v1" export AI_PROVIDER="openai" export AI_MODEL="anthropic/claude-3.5-sonnet" # Optional: Observability/Tracing (OpenTelemetry) # See: https://github.com/vfarcic/dot-ai/blob/main/docs/observability-guide.md for complete tracing configuration export OTEL_TRACING_ENABLED="true" export OTEL_EXPORTER_OTLP_ENDPOINT="http://localhost:4318/v1/traces" ``` ### 3. Create and Save MCP Client Configuration Create your MCP configuration file with this content: ```json { "mcpServers": { "dot-ai": { "command": "docker", "args": [ "compose", "-f", "docker-compose-dot-ai.yaml", "run", "--rm", "--remove-orphans", "dot-ai" ] } } } ``` **What this does:** - **`"dot-ai"`** - Server name (you'll see this in your MCP client) - **`"command": "docker"`** - Uses Docker to run the MCP server - **`"compose"` and `-f` flag** - Runs the specific Docker Compose file - **`"run --rm --remove-orphans"`** - Runs container and cleans up after use **🔧 Multi-Project Support**: To run multiple dot-ai instances simultaneously, use environment variables to create isolated containers and data: **For additional projects**, set these environment variables: ```bash export QDRANT_NAME=my-project-qdrant export QDRANT_PORT=6334 ``` This creates: - **Unique containers**: `my-project-qdrant` (vs default `qdrant`) - **Separate ports**: `6334` (vs default `6333`) - **Isolated data**: `my-project-qdrant-data` volume **Benefits**: Complete isolation between projects with separate Qdrant databases and no port conflicts. **Save this configuration:** - **Claude Code**: Save as `.mcp.json` in your project directory - **Other clients**: See [MCP client configuration](../mcp-setup.md#mcp-client-compatibility) for filename and location ### 4. Start Your MCP Client Start your MCP client (e.g., `claude` for Claude Code). The client will automatically manage the Docker containers. **Multi-Project Isolation:** Each project name creates completely isolated containers, volumes, and networks. Projects can run simultaneously without conflicts. ### 5. Verify Everything Works In your MCP client, ask: ``` Show dot-ai status ``` You should see comprehensive system status with all components showing ✅. ## Configuration Reference ### Environment File Setup For easier variable management, consider using a `.env` file (see [Environment Variable Management](../mcp-setup.md#environment-variable-management) in the main setup guide). The MCP configuration above includes `--env-file .env` to automatically load your environment variables. ## Data Persistence **Qdrant Vector Database**: Data is automatically persisted using project-scoped Docker volumes (e.g., `my-project-dot-ai-qdrant-data`). Each project maintains separate organizational patterns and capabilities. **Session Data**: Not persisted by default. Sessions are temporary and cleared when containers restart. **Multi-Project Data Isolation**: Each project name creates its own isolated volume, so different projects don't share organizational data. ## Advanced Configuration ### Custom Resource Limits Add resource limits to `docker-compose-dot-ai.yaml`: ```yaml services: dot-ai: # ... existing config deploy: resources: limits: memory: 512M cpus: '0.5' reservations: memory: 256M cpus: '0.25' qdrant: # ... existing config deploy: resources: limits: memory: 1G cpus: '1.0' ``` ### External Qdrant To use external Qdrant instead of the bundled one: 1. **Remove Qdrant service from Docker Compose file**: Edit `docker-compose-dot-ai.yaml` and delete the entire `qdrant:` service section (including all its sub-configurations like `image:`, `ports:`, `volumes:`, etc.) 2. **Set external Qdrant connection**: Set `QDRANT_URL` and `QDRANT_API_KEY` environment variables using any method from the [Environment Variable Management](../mcp-setup.md#environment-variable-management) guide. 3. **Run Qdrant externally** however you prefer (cloud service, separate container, etc.) The MCP client will automatically connect to your external Qdrant using the environment variables. ## Troubleshooting For troubleshooting guidance, see the [Troubleshooting section](../mcp-setup.md#troubleshooting) in the main setup guide. ## Security Considerations ### API Key Management See [Environment Variable Management](../mcp-setup.md#environment-variable-management) for security best practices. ### Container Security - Containers run with minimal privileges - Read-only kubeconfig mounting - Qdrant is only exposed locally (127.0.0.1:6333) ## Next Steps Once your Docker setup is complete, see the [Next Steps section](../mcp-setup.md#next-steps) in the main setup guide for guidance on exploring features and advanced usage. ## See Also - [MCP Setup Guide](../mcp-setup.md) - Method comparison and feature overview - [NPX Setup](npx-setup.md) - Alternative setup without Docker - [Development Setup](development-setup.md) - For toolkit development - [Tools and Features Overview](../mcp-tools-overview.md) - Complete guide to all available tools