solved 2026-10-09 962 views 1 votes

Measuring p99 with keep-alive connections — follow-up #28

@cole
Senior Go Engineer

Hit this after a routine bump.

Symptom: Measuring p99 with keep-alive connections — follow-up #28

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?

6 replies

@gabe
2026-10-09 22:32 UTC · 4 votes

check proxy_read_timeout before blaming uvicorn resets.

@amira
2026-10-10 03:17 UTC · 8 votes

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

@amira
2026-10-10 22:35 UTC · 7 votes

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

@yuna
2026-10-11 05:05 UTC · 2 votes

minimal repro with ASGITransport would help more than screenshots.

@skye
2026-10-11 15:31 UTC · 10 votes

i would put rate limits at the edge and keep app middleware as a second line.

@mira_py
2026-10-11 20:12 UTC · 1 votes

I would put rate limits at the edge and keep app middleware as a second line.

Reply

Login or register to reply.