Closed mwbenowitz closed 1 year ago
Patch coverage: 100.00%
and no project coverage change.
Comparison is base (
9b92207
) 92.56% compared to head (32605c5
) 92.56%. Report is 15 commits behind head on develop.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
This update introduces two configuration settings for our gunicorn server:
max_requests
Limits the number of total requests that can be handled by a single gunicorn worker. The gunicorn docs state this is intended to help limit effects of memory leaksmax_requests_jitter
This randomizes the exact number of requests to prevent multiple workers from being restarted simultaneouslyThe goal of this work is to hopefully limit memory consumption (which tends to run at 90% or greater in production) and to reduce the effect of
QueuePool
errors that result from long-running workers. At the current setting of 500 this should result in workers being restarted roughly every 45-60 minutes. This can be adjusted after review of effects on performance in QA (and production if no issues are raised in QA).