version: '3.8'

services:
  prometheus-api:
    build: .
    container_name: prometheus-api
    ports:
      - "7860:7860"
    environment:
      - PORT=7860
      - HOST=0.0.0.0
      - OMP_NUM_THREADS=4
      - MAX_INPUT_LENGTH=1500
      - ALLOWED_ORIGINS=*  
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    volumes:
      - model_cache:/home/appuser/.cache/huggingface
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:7860/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s

volumes:
  model_cache: