vLLM vllm/exceptions.py - the HTTP status carried by each admission-rejection exception

Fields

id

sources/w5e-vllm-exceptions-503-mapping

URL

https://github.com/vllm-project/vllm/blob/main/vllm/exceptions.py

Publisher

vllm-project

Kind

repo

Accessed

2026-10-05

notes

Read at v0.30.0 (release 2026-09-22, main commit 2026-10-05T04:19:19Z). QueueOverflowError and MaxQueuedTokensError both subclass GracefulHTTPError and both pass HTTPStatus.SERVICE_UNAVAILABLE (503) to super().__init__ alongside the retry-on-another-instance message. This file is the authoritative mapping from vLLM's internal admission exception to the wire status code; async_llm.py only documents the intent.

Notes

Read at v0.30.0 (release 2026-09-22, main commit 2026-10-05T04:19:19Z). QueueOverflowError and MaxQueuedTokensError both subclass GracefulHTTPError and both pass HTTPStatus.SERVICE_UNAVAILABLE (503) to super().__init__ alongside the retry-on-another-instance message. This file is the authoritative mapping from vLLM's internal admission exception to the wire status code; async_llm.py only documents the intent.

Cross-references

This record references no other records.