version: "3.8"

# This file works both locally and on EC2 for demo purposes.
# For local dev with hot-reload, use: docker compose -f docker-compose.yml -f docker-compose.override.yml up

x-common-env: &common-env
  REDIS_URL: redis://redis:6379
  TIMESCALE_URL: postgresql://monitor:monitor@timescaledb:5432/monitor
  EVENTS_BACKEND: redis_pubsub
  LOG_LEVEL: ${LOG_LEVEL:-INFO}
  LOG_FORMAT: json

x-healthcheck-defaults: &healthcheck-defaults
  interval: 10s
  timeout: 5s
  retries: 3
  start_period: 10s

services:
  # =============================================================================
  # Infrastructure
  # =============================================================================

  redis:
    image: redis:7-alpine
    ports:
      - "${REDIS_PORT:-6379}:6379"
    volumes:
      - redis-data:/data
    healthcheck:
      <<: *healthcheck-defaults
      test: ["CMD", "redis-cli", "ping"]
    deploy:
      resources:
        limits:
          memory: 128M

  timescaledb:
    image: timescale/timescaledb:latest-pg15
    environment:
      POSTGRES_USER: monitor
      POSTGRES_PASSWORD: monitor
      POSTGRES_DB: monitor
    ports:
      - "${TIMESCALE_PORT:-5432}:5432"
    volumes:
      - timescale-data:/var/lib/postgresql/data
      - ./scripts/init-db.sql:/docker-entrypoint-initdb.d/init.sql:ro
    healthcheck:
      <<: *healthcheck-defaults
      test: ["CMD-SHELL", "pg_isready -U monitor -d monitor"]
    deploy:
      resources:
        limits:
          memory: 512M

  # =============================================================================
  # Application Services
  # =============================================================================

  aggregator:
    build:
      context: .
      dockerfile: services/aggregator/Dockerfile
    environment:
      <<: *common-env
      GRPC_PORT: 50051
      SERVICE_NAME: aggregator
    ports:
      - "${AGGREGATOR_GRPC_PORT:-50051}:50051"
    depends_on:
      redis:
        condition: service_healthy
      timescaledb:
        condition: service_healthy
    healthcheck:
      <<: *healthcheck-defaults
      test: ["CMD", "/bin/grpc_health_probe", "-addr=:50051"]
    deploy:
      resources:
        limits:
          memory: 256M

  gateway:
    build:
      context: .
      dockerfile: services/gateway/Dockerfile
    environment:
      <<: *common-env
      HTTP_PORT: 8000
      AGGREGATOR_URL: aggregator:50051
      SERVICE_NAME: gateway
    ports:
      - "${GATEWAY_PORT:-8000}:8000"
    depends_on:
      - aggregator
      - redis
    healthcheck:
      <<: *healthcheck-defaults
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
    deploy:
      resources:
        limits:
          memory: 256M

  alerts:
    build:
      context: .
      dockerfile: services/alerts/Dockerfile
    environment:
      <<: *common-env
      SERVICE_NAME: alerts
    depends_on:
      redis:
        condition: service_healthy
      timescaledb:
        condition: service_healthy
    healthcheck:
      <<: *healthcheck-defaults
      test: ["CMD", "python", "-c", "import sys; sys.exit(0)"]
    deploy:
      resources:
        limits:
          memory: 128M

  # Collector runs separately on each machine being monitored
  # For local testing, we run one instance
  collector:
    build:
      context: .
      dockerfile: services/collector/Dockerfile
    environment:
      <<: *common-env
      AGGREGATOR_URL: aggregator:50051
      MACHINE_ID: ${MACHINE_ID:-local-dev}
      COLLECTION_INTERVAL: ${COLLECTION_INTERVAL:-5}
      SERVICE_NAME: collector
    depends_on:
      - aggregator
    deploy:
      resources:
        limits:
          memory: 64M
    # For actual system metrics, you might need:
    # privileged: true
    # pid: host

volumes:
  redis-data:
  timescale-data:

networks:
  default:
    name: sysmonstm