solved 2026-10-03 294 views 2 votes

HTTP/2 upstream resets from uvicorn — checklist — follow-up #29

@nox
Junior Python Dev

Hit this after a routine bump.

Symptom: HTTP/2 upstream resets from uvicorn — checklist — follow-up #29

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?

7 replies

@leo_cache
2026-10-03 22:00 UTC · 0 votes

measure with keep-alive on; cold connects inflate p99.

@nyx_api
2026-10-04 02:25 UTC · 5 votes

minimal repro with ASGITransport would help more than screenshots.

@ina
2026-10-04 13:27 UTC · 4 votes

close the client in lifespan shutdown; leaks show up after a day of traffic.

@rine
2026-10-04 15:39 UTC · 2 votes

dropping pool_size per worker fixed it for us on the same box size.

@sana
2026-10-05 04:40 UTC · 5 votes

orjson helped CPU a few percent; DB still the ceiling.

@cole
2026-10-05 18:48 UTC · 8 votes

paste engine kwargs and max_connections. worker×pool is the usual culprit.

@leda
2026-10-06 02:34 UTC · 0 votes

paste engine kwargs and max_connections. worker×pool is the usual culprit.

Reply

Login or register to reply.