Skip to content

vllm.entrypoints.serve.instrumentator.health

Functions:

health(raw_request) async

Health check.

Source code in vllm/entrypoints/serve/instrumentator/health.py 
22
23
24
25
26
27
28
29
30
31
32
33
@router.get("/health", response_class=Response)
async def health(raw_request: Request) -> Response:
    """Health check."""
    client = engine_client(raw_request)
    if client is None:
        # Render-only servers have no engine; they are always healthy.
        return Response(status_code=200)
    try:
        await client.check_health()
        return Response(status_code=200)
    except EngineDeadError:
        return Response(status_code=503)

health_generate(raw_request) async

GPU execution and EngineCore progress readiness check.

Source code in vllm/entrypoints/serve/instrumentator/health.py 
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
@router.get("/health_generate", response_class=Response)
async def health_generate(raw_request: Request) -> Response:
    """GPU execution and EngineCore progress readiness check."""
    client = engine_client(raw_request)
    if client is None:
        return Response(status_code=200)
    try:
        await client.check_health_gpu()
        return Response(status_code=200)
    except EngineDeadError:
        content = {"status": "not_ready", "reason": "engine_dead"}
    except EngineUnhealthyError as e:
        content = {
            "status": "not_ready",
            "reason": e.reason,
            "message": str(e),
            **e.details,
        }
    return JSONResponse(status_code=503, content=content)