phase11: /health/live + /health/ready endpoints + monitoring docs + Prometheus metrics docs

app/routes/health.py

@@ -1,8 +1,14 @@
-"""Health check endpoint — extended with DB, Redis, storage, worker checks."""
+"""Health check endpoints — liveness, readiness, and full health.
+
+/health/live   — process is alive (always 200 if running)
+/health/ready  — ready to serve requests (checks DB, Redis, storage)
+/api/v1/health — full health with all checks (backward compatible)
+"""
 
 from __future__ import annotations
 
 from fastapi import APIRouter
+from pydantic import BaseModel
 
 from app.core.monitoring import get_health_status
 from app.schemas.common import HealthResponse
@@ -10,9 +16,57 @@ from app.schemas.common import HealthResponse
 router = APIRouter(tags=["health"])
 
 
+class LiveResponse(BaseModel):
+    status: str = "alive"
+
+
+class ReadyResponse(BaseModel):
+    status: str  # "ready" or "not_ready"
+    checks: dict[str, str] = {}
+
+
+@router.get("/health/live")
+async def health_live() -> LiveResponse:
+    """Liveness probe — process is alive.
+
+    Returns 200 if the process is running.
+    Used by Kubernetes/Coolify for restart decisions.
+    """
+    return LiveResponse(status="alive")
+
+
+@router.get("/health/ready")
+async def health_ready() -> ReadyResponse:
+    """Readiness probe — ready to serve requests.
+
+    Checks:
+    - PostgreSQL connection
+    - Redis connection
+    - Storage backend
+    - Worker heartbeat (if available)
+
+    Returns 200 if all checks pass, 503 if any fail.
+    Used by load balancer to route traffic.
+    """
+    health = await get_health_status()
+    checks = {}
+    all_ready = True
+
+    for check_name, check_val in (health.checks or {}).items():
+        status = "ok" if check_val == "ok" or check_val is True else "fail"
+        checks[check_name] = status
+        if status != "ok":
+            all_ready = False
+
+    return ReadyResponse(
+        status="ready" if all_ready else "not_ready",
+        checks=checks,
+    )
+
+
 @router.get("/api/v1/health", response_model=HealthResponse)
 async def health():
-    """Health check — no auth required.
+    """Full health check — no auth required.
 
     Returns status (healthy/degraded) and individual checks for
     database, redis, storage, and worker.