@piotr
2026-10-12 15:59 UTC · 1 votes
dropping pool_size per worker fixed it for us on the same box size.
Hit this after a routine bump.
Symptom: orjson Response vs JSONResponse in practice — follow-up #12
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?
dropping pool_size per worker fixed it for us on the same box size.
close the client in lifespan shutdown; leaks show up after a day of traffic.
measure with keep-alive on; cold connects inflate p99.
minimal repro with ASGITransport would help more than screenshots.
I would put rate limits at the edge and keep app middleware as a second line.
measure with keep-alive on; cold connects inflate p99.
that traceback looks like Engine created per request — move it to lifespan.
check proxy_read_timeout before blaming uvicorn resets.