zhifu-platform/compose.runtime.yaml

name: zhifu-runtime
x-runtime: &runtime
  image: ${ZHIFU_IMAGE:-zhifu-platform:local}
  env_file: ${ZHIFU_RUNTIME_ENV:?Specify an absolute runtime environment file path}
  init: true
  read_only: true
  tmpfs: ["/app/.work:uid=10001,gid=10001,mode=0700"]
  cap_drop: [ALL]
  security_opt: ["no-new-privileges:true"]
  stop_grace_period: 75s
  restart: unless-stopped
  pids_limit: 256
  mem_limit: 1g
  cpus: 2
  logging:
    driver: json-file
    options: {max-size: "10m", max-file: "3"}
  networks: [dependencies]
services:
  api:
    <<: *runtime
    environment:
      DEEPSEEK_API_KEY: ${DEEPSEEK_API_KEY:-}
    ports: ["127.0.0.1:${ZHIFU_API_PORT:-28000}:8000"]
    healthcheck:
      test: [CMD, python, -c, "import httpx; httpx.get('http://127.0.0.1:8000/health/ready', timeout=4).raise_for_status()"]
      interval: 10s
      timeout: 5s
      retries: 3
      start_period: 30s
  oms:
    <<: *runtime
    command: [uvicorn, "zhifu.oms.api:create_app", --factory, --host, "0.0.0.0", --port, "8100", --timeout-graceful-shutdown, "60"]
    ports: ["127.0.0.1:${ZHIFU_OMS_PORT:-28100}:8100"]
    healthcheck:
      test: [CMD, python, -c, "import httpx; httpx.get('http://127.0.0.1:8100/health', timeout=4).raise_for_status()"]
      interval: 10s
      timeout: 5s
      retries: 3
      start_period: 30s
  operations:
    <<: *runtime
    command: [uvicorn, "zhifu.monitoring.api:app", --host, "0.0.0.0", --port, "8300"]
    ports: ["127.0.0.1:28300:8300"]
  agent:
    <<: *runtime
    profiles: [workers]
    environment:
      DEEPSEEK_API_KEY: ${DEEPSEEK_API_KEY:-}
    command: [python, -m, zhifu.operations.worker, agent]
  memory:
    <<: *runtime
    profiles: [workers]
    environment:
      DEEPSEEK_API_KEY: ${DEEPSEEK_API_KEY:-}
    command: [python, -m, zhifu.operations.worker, memory]
  fulfillment:
    <<: *runtime
    profiles: [workers]
    command: [python, -m, zhifu.operations.worker, fulfillment]
  knowledge:
    <<: *runtime
    profiles: [workers]
    command: [python, -m, zhifu.operations.worker, knowledge]
  channel-delivery:
    <<: *runtime
    profiles: [workers]
    command: [python, -m, zhifu.operations.worker, channel-delivery]
  migrate:
    <<: *runtime
    profiles: [release]
    restart: "no"
    command: [alembic, upgrade, head]
  migrate-oms:
    <<: *runtime
    profiles: [release]
    restart: "no"
    command: [alembic, -c, oms-alembic.ini, upgrade, head]
  release-check:
    <<: *runtime
    profiles: [release]
    environment:
      DEEPSEEK_API_KEY: ${DEEPSEEK_API_KEY:-}
    restart: "no"
    command: [python, -m, zhifu.operations.release_check]
  inference:
    <<: *runtime
    image: ${ZHIFU_INFERENCE_IMAGE:-zhifu-inference:local}
    profiles: [inference]
    command: [uvicorn, "zhifu.modules.knowledge.inference:app", --host, "0.0.0.0", --port, "8200", --timeout-graceful-shutdown, "60"]
    environment:
      KNOWLEDGE_MODEL_CACHE: /models
      HF_HUB_OFFLINE: "1"
      HF_HOME: /app/.work/huggingface
    volumes: ["${ZHIFU_MODEL_PATH:?Specify the BGE model cache path}:/models:ro"]
    mem_limit: 6g
    cpus: 4
    ports: ["127.0.0.1:28200:8200"]
    healthcheck:
      test: [CMD, python, -c, "import httpx; httpx.get('http://127.0.0.1:8200/health/ready', timeout=4).raise_for_status()"]
      interval: 10s
      timeout: 5s
      retries: 6
      start_period: 120s
networks:
  dependencies:
    external: true
    name: ${ZHIFU_DEPENDENCY_NETWORK:-zhifu-platform_default}