Skip to content

Docker Compose

Deploy the full MPL stack locally with Docker Compose. This includes the proxy, a demo MCP server, and optional monitoring.

Quick Deploy

git clone https://github.com/Skelf-Research/mpl.git
cd mpl

# Start core services (proxy + demo server)
docker compose up -d

# Verify
curl http://localhost:9443/health

Services

The docker-compose.yaml includes:

Service Port Description
mpl-proxy 9443, 9100 MPL sidecar proxy with metrics
demo-server 8080 Demo MCP server (echo server for testing)
prometheus 9090 Metrics collection (optional, monitoring profile)
grafana 3000 Dashboards (optional, monitoring profile)

Core Stack

The default docker compose up starts the proxy and demo server:

services:
  mpl-proxy:
    build: .
    ports:
      - "9443:9443"   # Proxy endpoint
      - "9100:9100"   # Prometheus metrics
    volumes:
      - ./registry:/app/registry:ro
      - ./mpl-config.yaml:/app/mpl-config.yaml:ro
    environment:
      - RUST_LOG=info

  demo-server:
    image: python:3.11-slim
    working_dir: /app
    volumes:
      - ./examples/demo-server:/app
    command: python server.py
    ports:
      - "8080:8080"

With Monitoring

To include Prometheus and Grafana:

docker compose --profile monitoring up -d

This adds:

  • Prometheus at http://localhost:9090 — scrapes MPL metrics every 15s
  • Grafana at http://localhost:3000 — visualize metrics (login: admin/admin)

Configuration

The proxy reads from mpl-config.yaml:

transport:
  listen: "0.0.0.0:9443"
  upstream: "demo-server:8080"
  protocol: http

mpl:
  registry: "./registry"
  mode: transparent          # transparent | strict
  required_profile: qom-basic
  enforce_schema: true
  enforce_assertions: true

observability:
  metrics_port: 9100
  metrics_format: prometheus
  logs: stdout
  log_format: json
  log_level: info

Switching to Strict Mode

Change mode: transparent to mode: strict and restart to enforce validation. In strict mode, invalid payloads are rejected with typed errors.

Testing the Stack

Once running, test validation through the proxy:

# Health check
curl http://localhost:9443/health

# Check capabilities
curl http://localhost:9443/capabilities

# Validate a payload
curl -X POST http://localhost:9443/validate \
  -H "Content-Type: application/json" \
  -d '{
    "stype": "org.calendar.Event.v1",
    "payload": {
      "title": "Meeting",
      "start": "2025-01-15T10:00:00Z",
      "end": "2025-01-15T11:00:00Z"
    }
  }'

Customizing

Point to Your Own Server

Replace the demo server with your MCP/A2A server by editing mpl-config.yaml:

transport:
  upstream: "your-mcp-server:8080"

Or override via environment variable:

MPL_UPSTREAM=http://your-server:8080 docker compose up -d

Add Your Own STypes

Mount your custom registry alongside the default:

volumes:
  - ./my-registry:/app/registry:ro

Resource Limits

For production-like testing, add resource constraints:

services:
  mpl-proxy:
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 256M

Stopping

# Stop all services
docker compose down

# Stop and remove volumes
docker compose down -v

Next Steps