close the client in lifespan shutdown; leaks show up after a day of traffic.
Deadlocks with nested begin() in SQLAlchemy 2 — follow-up #18
@gabe
Senior Python Dev
Hit this after a routine bump.
Symptom: Deadlocks with nested begin() in SQLAlchemy 2 — follow-up #18
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
@cleo_dev29
2026-10-17 03:59 UTC · 0 votes
@leda
2026-10-17 15:03 UTC · 0 votes
paste engine kwargs and max_connections. worker×pool is the usual culprit.
@raul
2026-10-18 05:08 UTC · 5 votes
orjson helped CPU a few percent; DB still the ceiling.
@cole
2026-10-18 23:40 UTC · 10 votes
check proxy_read_timeout before blaming uvicorn resets.
@elias
2026-10-19 16:44 UTC · 6 votes
paste engine kwargs and max_connections. worker×pool is the usual culprit.
@tomo
2026-10-19 21:40 UTC · 8 votes
dropping pool_size per worker fixed it for us on the same box size.
@cole
2026-10-20 02:46 UTC · 9 votes
minimal repro with ASGITransport would help more than screenshots.
@ina
2026-10-20 14:08 UTC · 3 votes
that traceback looks like Engine created per request — move it to lifespan.
@brick
2026-10-21 04:56 UTC · 0 votes
that traceback looks like engine created per request — move it to lifespan.