that traceback looks like Engine created per request — move it to lifespan.
When background tasks should be a queue instead — follow-up #20
@omar
Tech Lead
Hit this after a routine bump.
Symptom: When background tasks should be a queue instead — follow-up #20
TimeoutError / pool exhausted / upstream reset
(versions in next reply if needed)
Env: Python 3.12, FastAPI recent, uvicorn workers≥2.
What should I check first?
9 replies
@ina
2026-11-11 00:48 UTC · 0 votes
@ina
2026-11-11 13:54 UTC · 0 votes
minimal repro with ASGITransport would help more than screenshots.
@nik
2026-11-11 16:19 UTC · 5 votes
I would put rate limits at the edge and keep app middleware as a second line.
@nyx_api
2026-11-12 06:24 UTC · 10 votes
marking solved after we pinned the dependency and re-smoked /docs.
@gabe
2026-11-12 23:25 UTC · 6 votes
dropping pool_size per worker fixed it for us on the same box size.
@leda
2026-11-13 12:26 UTC · 9 votes
measure with keep-alive on; cold connects inflate p99.
@seth
2026-11-14 03:35 UTC · 2 votes
orjson helped CPU a few percent; DB still the ceiling.
@raul
2026-11-14 05:19 UTC · 8 votes
I would put rate limits at the edge and keep app middleware as a second line.
@cleo_dev29
2026-11-14 16:09 UTC · 1 votes
that traceback looks like Engine created per request — move it to lifespan.