Skip to content

vllm.entrypoints.scale_out.token_in_token_out.api_router

Functions:

abort_requests(raw_request) async

Abort one or more requests. To be used in a Disaggregated Everything setup.

Source code in vllm/entrypoints/scale_out/token_in_token_out/api_router.py 
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
async def abort_requests(raw_request: Request):
    """Abort one or more requests. To be used in a
    Disaggregated Everything setup.
    """
    try:
        body = await raw_request.json()
    except json.JSONDecodeError as e:
        raise HTTPException(
            status_code=HTTPStatus.BAD_REQUEST.value,
            detail=f"JSON decode error: {e}",
        ) from e
    # Validate the parsed body instead of declaring it as a FastAPI body param:
    # FastAPI's strict content-type check would then reject clients that send
    # no `Content-Type`, which the legacy `/abort_requests` path still allows.
    try:
        request = AbortRequestsRequest.model_validate(body)
    except ValidationError as e:
        raise RequestValidationError(
            errors=[{**err, "loc": ("body", *err["loc"])} for err in e.errors()]
        ) from e
    # Abort requests in background
    asyncio.create_task(engine_client(raw_request).abort(request.request_ids))
    return Response(status_code=200)