Gunicorn worker timeout

[furgerf]

Is your feature request related to a problem? Please describe.
I noticed that Sentry metrics were no longer exported and saw this timeout - presumably while retrieving data from Sentry, though it's not quite clear what exactly hits the timeout:

[2024-03-19 15:49:25 +0000] [292] [INFO] exporter: cleaning registry collectors...
[2024-03-19 15:49:25 +0000] [292] [INFO] exporter: cleaning registry collectors...
[2024-03-19 15:49:26 +0000] [292] [INFO] metadata: sentry organization: <org>
[2024-03-19 15:49:26 +0000] [292] [INFO] metadata: no projects specified, loading from API
[2024-03-19 15:49:28 +0000] [292] [INFO] metadata: projects loaded from API: 48
[2024-03-19 15:50:12 +0000] [292] [INFO] collector: loading projects issues
[2024-03-19 15:50:25 +0000] [292] [INFO] metadata: sentry organization: <org>
[2024-03-19 15:50:25 +0000] [292] [INFO] metadata: no projects specified, loading from API
[2024-03-19 15:50:26 +0000] [1] [CRITICAL] WORKER TIMEOUT (pid:292)
[2024-03-19 15:50:26 +0000] [292] [INFO] Worker exiting (pid: 292)
[2024-03-19 15:50:26 +0000] [431] [INFO] Booting worker with pid: 431

Describe the solution you'd like
I worked around it by increasing the default timeout of 30s with GUNICORN_CMD_ARGS but ideally, there would be a parameter for that in this project 🙂 It may also be helpful to document this.

Describe alternatives you've considered
I'm not quite sure how this is intended to work, e.g. it seems to export data every 5 mins and there are explicitly 4 gunicorn workers. I'm not sure what the 4 workers are for, are they for the case where one export takes more than 5 mins? With the default timeout, that isn't possible. Maybe there's another way to improve this, e.g. retrieve the different issue time horizons on separate workers.

Additional context
The Sentry instance is self-hosted and possibly our number of projects is rather high (48).