docker-compose.multi-agent.yml

version: '3.8'

networks:
  agent-network:
    driver: bridge
    ipam:
      config:
        - subnet: 172.20.0.0/16

volumes:
  redis-data:
  prometheus-data:
  grafana-data:
  workspace:
  artifacts:

services:
  # Core Infrastructure
  redis:
    image: redis:7-alpine
    container_name: agent-redis
    ports:
      - "6379:6379"
    volumes:
      - redis-data:/data
    networks:
      - agent-network
    command: redis-server --appendonly yes --maxmemory 512mb --maxmemory-policy allkeys-lru
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 10s
      timeout: 3s
      retries: 3

  # Main Orchestrator
  orchestrator:
    build: ./orchestrator
    container_name: agent-orchestrator
    ports:
      - "8000:8000"
    environment:
      - ORCHESTRATOR_MODE=${ORCHESTRATION_MODE:-hybrid}
      - REDIS_HOST=redis
      - MAX_PARALLEL_AGENTS=4
      - LOG_LEVEL=INFO
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
      - ./workspace:/workspace
      - ./artifacts:/artifacts
      - ./config:/config
    networks:
      - agent-network
    depends_on:
      redis:
        condition: service_healthy
    deploy:
      resources:
        limits:
          memory: 1G
          cpus: '1.0'

  # Claude Agent - Architecture & Design
  claude-architect:
    image: anthropic/claude-agent:latest
    container_name: claude-architect
    environment:
      - AGENT_ROLE=architect
      - CLAUDE_API_KEY_FILE=/run/secrets/claude_api_key
      - MODEL=claude-3-opus-20240229
      - FOCUS_AREAS=architecture,design,patterns,security
      - REDIS_HOST=redis
      - MAX_TOKENS=4096
    volumes:
      - ./workspace:/workspace
      - ./artifacts:/artifacts
    networks:
      - agent-network
    secrets:
      - claude_api_key
    deploy:
      replicas: 1
      resources:
        limits:
          memory: 2G
          cpus: '1.0'
    profiles:
      - architects

  # Gemini Agent - Implementation
  gemini-developer:
    image: google/gemini-agent:latest
    container_name: gemini-developer
    environment:
      - AGENT_ROLE=developer
      - GEMINI_API_KEY_FILE=/run/secrets/gemini_api_key
      - MODEL=gemini-2.5-pro
      - FOCUS_AREAS=implementation,optimization,algorithms
      - REDIS_HOST=redis
      - CONTEXT_WINDOW=1000000
    volumes:
      - ./workspace:/workspace
      - ./artifacts:/artifacts
    networks:
      - agent-network
    secrets:
      - gemini_api_key
    deploy:
      replicas: 1
      resources:
        limits:
          memory: 2G
          cpus: '1.0'
    profiles:
      - developers

  # Claude Agent - Testing
  claude-tester:
    image: anthropic/claude-agent:latest
    container_name: claude-tester
    environment:
      - AGENT_ROLE=tester
      - CLAUDE_API_KEY_FILE=/run/secrets/claude_api_key
      - MODEL=claude-3-sonnet-20240229
      - FOCUS_AREAS=testing,verification,quality
      - REDIS_HOST=redis
    volumes:
      - ./workspace:/workspace
      - ./artifacts:/artifacts
    networks:
      - agent-network
    secrets:
      - claude_api_key
    deploy:
      replicas: 1
      resources:
        limits:
          memory: 1G
          cpus: '0.5'
    profiles:
      - testers

  # Codeium Agent - Refactoring
  codeium-refactorer:
    image: codeium/agent:latest
    container_name: codeium-refactorer
    environment:
      - AGENT_ROLE=refactorer
      - CODEIUM_API_KEY_FILE=/run/secrets/codeium_api_key
      - FOCUS_AREAS=refactoring,cleanup,performance
      - REDIS_HOST=redis
    volumes:
      - ./workspace:/workspace
      - ./artifacts:/artifacts
    networks:
      - agent-network
    secrets:
      - codeium_api_key
    deploy:
      replicas: 1
      resources:
        limits:
          memory: 1G
          cpus: '0.5'
    profiles:
      - refactorers

  # Monitoring Stack
  prometheus:
    image: prom/prometheus:latest
    container_name: agent-prometheus
    volumes:
      - ./monitoring/prometheus.yml:/etc/prometheus/prometheus.yml
      - prometheus-data:/prometheus
    ports:
      - "9090:9090"
    networks:
      - agent-network
    command:
      - '--config.file=/etc/prometheus/prometheus.yml'
      - '--storage.tsdb.path=/prometheus'
      - '--web.console.libraries=/usr/share/prometheus/console_libraries'
      - '--web.console.templates=/usr/share/prometheus/consoles'
    profiles:
      - monitoring

  grafana:
    image: grafana/grafana:latest
    container_name: agent-grafana
    environment:
      - GF_SECURITY_ADMIN_USER=admin
      - GF_SECURITY_ADMIN_PASSWORD=admin
      - GF_INSTALL_PLUGINS=redis-datasource
    volumes:
      - grafana-data:/var/lib/grafana
      - ./monitoring/grafana/dashboards:/etc/grafana/provisioning/dashboards
      - ./monitoring/grafana/datasources:/etc/grafana/provisioning/datasources
    ports:
      - "3001:3000"
    networks:
      - agent-network
    depends_on:
      - prometheus
    profiles:
      - monitoring

  # Container metrics
  cadvisor:
    image: gcr.io/cadvisor/cadvisor:latest
    container_name: agent-cadvisor
    volumes:
      - /:/rootfs:ro
      - /var/run:/var/run:ro
      - /sys:/sys:ro
      - /var/lib/docker:/var/lib/docker:ro
      - /dev/disk/:/dev/disk:ro
    ports:
      - "8080:8080"
    networks:
      - agent-network
    privileged: true
    devices:
      - /dev/kmsg
    profiles:
      - monitoring

  # Resource Manager
  resource-manager:
    build: ./resource-manager
    container_name: agent-resource-manager
    environment:
      - MAX_TOTAL_MEMORY=16G
      - MAX_TOTAL_CPU=8
      - REDIS_HOST=redis
      - CHECK_INTERVAL=30
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    networks:
      - agent-network
    depends_on:
      - redis
      - orchestrator
    deploy:
      resources:
        limits:
          memory: 512M
          cpus: '0.5'
    profiles:
      - management

  # Persistence Service for Agent Outputs
  persistence-service:
    build: ./persistence
    container_name: agent-persistence
    ports:
      - "5001:5000"
    environment:
      - WORKSPACE_DIR=/workspace
      - SPECS_DIR=/specs
      - OUTPUTS_DIR=/outputs
      - VALIDATE_SPECS=true
      - AUTO_COMMIT=true
      - TDD_MODE=enforced
      - REDIS_HOST=redis
    volumes:
      - ./workspace:/workspace
      - ./specs:/specs
      - ./outputs:/outputs
      - ./artifacts:/artifacts
    networks:
      - agent-network
    depends_on:
      redis:
        condition: service_healthy
    deploy:
      resources:
        limits:
          memory: 512M
          cpus: '0.5'
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:5000/status"]
      interval: 10s
      timeout: 3s
      retries: 3

secrets:
  claude_api_key:
    file: ~/.secrets/claude_api_key
  gemini_api_key:
    file: ~/.secrets/gemini_api_key
  codeium_api_key:
    file: ~/.secrets/codeium_api_key

# Usage Examples:
# 
# Start core services only:
# docker-compose -f docker-compose.multi-agent.yml up orchestrator redis
#
# Start with monitoring:
# docker-compose -f docker-compose.multi-agent.yml --profile monitoring up
#
# Start all architect agents:
# docker-compose -f docker-compose.multi-agent.yml --profile architects up
#
# Start everything:
# docker-compose -f docker-compose.multi-agent.yml --profile architects --profile developers --profile testers --profile monitoring --profile management up
#
# Scale specific agents:
# docker-compose -f docker-compose.multi-agent.yml up --scale gemini-developer=3