Sentry-python: Sentry not logging Gunicorn timeout errors

[hyperknot]

We have quite a standard gunicorn setup with the default sync worker with the default 1 thread, running a Python app.

Gunicorn is started with this command line:

gunicorn \
    --workers=16 \
    --timeout=120 \
    -b 127.0.0.1:8082 \
    --paste production.ini \
    2>> api-err.log 1>> api-out.log

The Python Sentry integration is loaded with the following:

import sentry_sdk
from sentry_sdk.integrations.pyramid import PyramidIntegration

sentry_sdk.init(
    dsn="https://[email protected]/123",
    integrations=[PyramidIntegration()],
    release='1.2.3'
)

config = Configurator(settings=settings)
...

The issue is that whenever a Gunicorn worker timeout occurs, it is not logged to Sentry.
The only way we can see that there was a timeout is when we see "[CRITICAL] WORKER TIMEOUT" in api-err.log and the matching error code 499 lines in nginx.log.

It'd be important to log these timeout events to Sentry.

[untitaker]

Could you write a python script that contains the same thing as your gunicorn executable, but calls sentry_sdk.init before running gunicorn? Trying to confirm something

[hyperknot]

This is the gunicorn executable:

```

!/opt/conda/bin/python

-- coding: utf-8 --

import re
import sys

from gunicorn.app.wsgiapp import run

if __name__ == '__main__':
sys.argv[0] = re.sub(r'(-script.pyw?|.exe)?$', '', sys.argv[0])
sys.exit(run())
````

[untitaker]

Yes, could you edit it to call sentry_sdk.init() (with your DSN) before it calls run()? I believe this could "fix" the problem already.

[darwinyip]

Bump on the issue. Similar configuration but with gunicorn + falcon + sentry sdk.