Config for running wsgi app with memory leak

[pinguin999]

Hi,
we have a nice django app we run via wsgi (with 3 workers) and we see a linear ram increase.

The memory leak is not in granian we also have the same problem with gunicorn.

1. What settings would you recommend for running this app with granian?
2. If you have any tips, on how to find the leak, it would also help a lot.

Thanks
Kolja

[gi0baro]

--workers-lifetime to 1800-3600

[pinguin999]

Sounds good, but is there any kind of logic that would prevent my 3 workers from being restarted at the same time?

[gi0baro]

You can tune the --respawn-interval option. See also #372 (comment)

[pinguin999]

Thanks @gi0baro for your help.

The Problem solved itself with the last Sentry-SDK update.....
https://github.com/getsentry/sentry-python/releases/tag/2.24.0 -> Fix memory leak by not piling up breadcrumbs forever in Spark workers. (getsentry/sentry-python#4167) by @antonpirker

We will keep these changes ;)