docker-compose.yml

# ============================================================================
# ModelServe — Docker Compose (Local Development)
# ============================================================================
# Complete local development stack with all services.

# ─────────────────────────────────────────────────────────────
#  Networks & Volumes
# ─────────────────────────────────────────────────────────────

networks:
  modelserve:
    driver: bridge

volumes:
  postgres_data:
  mlflow_artifacts:
  grafana_data:

services:

  # ── 1. PostgreSQL ─────────────────────────────────────────
  postgres:
    image: postgres:15-alpine
    container_name: modelserve-postgres
    restart: unless-stopped
    networks: [modelserve]
    environment:
      POSTGRES_USER:     ${POSTGRES_USER:-mlflow}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-mlflow}
      POSTGRES_DB:       ${POSTGRES_DB:-mlflow}
    ports:
      - "5432:5432"
    volumes:
      - postgres_data:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-mlflow}"]
      interval: 5s
      timeout: 5s
      retries: 10

  # ── 2. Redis (Feast online store) ─────────────────────────
  redis:
    image: redis:7-alpine
    container_name: modelserve-redis
    restart: unless-stopped
    networks: [modelserve]
    ports:
      - "6379:6379"
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 3s
      retries: 5

  # ── 3. MLflow Tracking Server ─────────────────────────────
  mlflow:
    image: ghcr.io/mlflow/mlflow:v2.12.2
    container_name: modelserve-mlflow
    restart: unless-stopped
    networks: [modelserve]
    depends_on:
      postgres:
        condition: service_healthy
    ports:
      - "5000:5000"
    environment:
      MLFLOW_TRACKING_URI: postgresql+psycopg2://${POSTGRES_USER:-mlflow}:${POSTGRES_PASSWORD:-mlflow}@postgres:5432/${POSTGRES_DB:-mlflow}
      AWS_ACCESS_KEY_ID:     ${AWS_ACCESS_KEY_ID:-}
      AWS_SECRET_ACCESS_KEY: ${AWS_SECRET_ACCESS_KEY:-}
      AWS_DEFAULT_REGION:    ${AWS_DEFAULT_REGION:-ap-southeast-1}
    volumes:
      - mlflow_artifacts:/mlflow/artifacts
    command: >
      sh -c "pip install psycopg2-binary && mlflow server
      --backend-store-uri postgresql+psycopg2://${POSTGRES_USER:-mlflow}:${POSTGRES_PASSWORD:-mlflow}@postgres:5432/${POSTGRES_DB:-mlflow}
      --artifacts-destination /mlflow/artifacts
      --host 0.0.0.0
      --port 5000"
    healthcheck:
      test: ["CMD", "curl", "-sf", "http://localhost:5000/"]
      interval: 10s
      timeout: 5s
      retries: 10

  # ── 4. FastAPI Inference Service ──────────────────────────
  fastapi:
    build:
      context: .
      dockerfile: Dockerfile
    container_name: modelserve-fastapi
    restart: unless-stopped
    networks: [modelserve]
    depends_on:
      mlflow:
        condition: service_started
      redis:
        condition: service_healthy
    ports:
      - "8000:8000"
    environment:
      MLFLOW_TRACKING_URI: http://mlflow:5000
      MODEL_NAME:          ${MODEL_NAME:-FraudDetector}
      MODEL_STAGE:         ${MODEL_STAGE:-Production}
      REDIS_HOST:          redis
      REDIS_PORT:          6379
      FEAST_REPO_PATH:     /app/feast_repo
      AWS_ACCESS_KEY_ID:     ${AWS_ACCESS_KEY_ID:-}
      AWS_SECRET_ACCESS_KEY: ${AWS_SECRET_ACCESS_KEY:-}
      AWS_DEFAULT_REGION:    ${AWS_DEFAULT_REGION:-ap-southeast-1}
    volumes:
      - ./feast_repo:/app/feast_repo:rw
      - ./training:/app/training:ro
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 10s
      timeout: 5s
      retries: 5

  # ── 5. Prometheus ─────────────────────────────────────────
  prometheus:
    image: prom/prometheus:v2.52.0
    container_name: modelserve-prometheus
    restart: unless-stopped
    networks: [modelserve]
    ports:
      - "9090:9090"
    volumes:
      - ./monitoring/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
      - ./monitoring/prometheus/alerts.yml:/etc/prometheus/alerts.yml:ro
    command:
      - "--config.file=/etc/prometheus/prometheus.yml"
      - "--storage.tsdb.path=/prometheus"
      - "--web.enable-lifecycle"
    healthcheck:
      test: ["CMD", "wget", "-qO-", "http://localhost:9090/-/healthy"]
      interval: 10s
      timeout: 5s
      retries: 5

  # ── 6. Grafana ────────────────────────────────────────────
  grafana:
    image: grafana/grafana:10.4.3
    container_name: modelserve-grafana
    restart: unless-stopped
    networks: [modelserve]
    depends_on:
      - prometheus
    ports:
      - "3001:3000"
    environment:
      GF_SECURITY_ADMIN_PASSWORD: ${GF_SECURITY_ADMIN_PASSWORD:-admin}
      GF_USERS_ALLOW_SIGN_UP:     "false"
      GF_SERVER_SERVE_FROM_SUB_PATH: "true"
    volumes:
      - grafana_data:/var/lib/grafana
      - ./monitoring/grafana/provisioning:/etc/grafana/provisioning:ro
      - ./monitoring/grafana/dashboards:/var/lib/grafana/dashboards:ro