feat(backend): add unauthenticated /health liveness endpoint

Add GET /health mounted outside the /api prefix so it bypasses
per-route auth and is reachable as a plain liveness/readiness probe.
Returns {status, version, db_ok}, degrading to db_ok=false with a 200
(not a raised exception) on database failure so the probe stays useful
even when the DB is unreachable. Add a Docker HEALTHCHECK that curls it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GTnuEgR78cBk32bLNKar3z

Dockerfile

@@ -46,4 +46,9 @@ ENV PATH="/app/.venv/bin:$PATH"
 EXPOSE 7892
 VOLUME ["/app/config", "/app/data"]
 
+# Liveness probe against the unauthenticated /health route; give the app
+# room to run migrations on first boot before failures count.
+HEALTHCHECK --interval=30s --timeout=5s --start-period=30s --retries=3 \
+    CMD wget -qO- http://localhost:7892/health || exit 1
+
 ENTRYPOINT ["tini", "-g", "--", "/entrypoint.sh"]

backend/src/main.py

@@ -11,6 +11,7 @@ from fastapi.staticfiles import StaticFiles
 from fastapi.templating import Jinja2Templates
 
 from module.api import v1
+from module.api.health import router as health_router
 from module.conf import VERSION, settings, setup_logger
 from module.core import AppContext
 from module.mcp import create_mcp_app
@@ -65,6 +66,10 @@ def create_app() -> FastAPI:
     # mount routers
     app.include_router(v1, prefix="/api")
 
+    # unauthenticated liveness probe, mounted at the app root (not /api) so
+    # it stays reachable without auth for container/orchestrator health checks
+    app.include_router(health_router)
+
     # mount MCP server (SSE transport for LLM tool integration)
     app.mount("/mcp", create_mcp_app(ctx))
 

backend/src/module/api/health.py

@@ -0,0 +1,29 @@
+"""无鉴权的存活探针（liveness probe），供容器/编排系统健康检查使用。
+
+与 program.py 中的 /status 不同，本路由不挂在 /api/v1 前缀下、也不依赖
+get_current_user，因此可以直接被 Docker HEALTHCHECK 等外部探针访问。
+"""
+
+import logging
+
+from fastapi import APIRouter
+from sqlalchemy import text
+
+from module.conf import VERSION
+from module.database import Database
+
+logger = logging.getLogger(__name__)
+router = APIRouter(tags=["health"])
+
+
+@router.get("/health")
+async def health() -> dict:
+    """存活探针：始终返回 200，db_ok 反映一次简单查询是否成功。"""
+    try:
+        async with Database() as db:
+            await db.session.execute(text("SELECT 1"))
+        db_ok = True
+    except Exception as e:
+        logger.warning(f"[Health] DB check failed: {e}")
+        db_ok = False
+    return {"status": "ok", "version": VERSION, "db_ok": db_ok}

backend/src/test/test_api_health.py

@@ -0,0 +1,60 @@
+"""Tests for the unauthenticated /health liveness probe."""
+
+from unittest.mock import patch
+
+import pytest
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from module.api.health import router as health_router
+
+# ---------------------------------------------------------------------------
+# Fixtures
+# ---------------------------------------------------------------------------
+
+
+@pytest.fixture
+def app():
+    """App with only the health router mounted, at the root (no /api prefix)."""
+    app = FastAPI()
+    app.include_router(health_router)
+    return app
+
+
+@pytest.fixture
+def client(app):
+    """Plain TestClient — no auth override, matching the probe's real usage."""
+    return TestClient(app)
+
+
+# ---------------------------------------------------------------------------
+# GET /health
+# ---------------------------------------------------------------------------
+
+
+class TestHealth:
+    def test_health_no_auth_required_returns_200(self, client):
+        """GET /health succeeds without any auth dependency/header."""
+        response = client.get("/health")
+        assert response.status_code == 200
+
+    def test_health_reports_version_and_db_ok(self, client):
+        """GET /health returns status/version/db_ok, DB reachable -> db_ok True."""
+        with patch("module.api.health.VERSION", "3.3.0-test"):
+            response = client.get("/health")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "ok"
+        assert data["version"] == "3.3.0-test"
+        assert data["db_ok"] is True
+
+    def test_health_db_failure_reports_db_ok_false(self, client):
+        """A DB query failure is caught: still 200, but db_ok is False."""
+        with patch("module.api.health.Database", side_effect=RuntimeError("db down")):
+            response = client.get("/health")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "ok"
+        assert data["db_ok"] is False