database crashes regularly and is restarted

[zven]

According to our observation and logs, the database is crashing "regularly" on prod and staging (albeit more often here) and is automatically restarted via init.sql.

@mpfeil hast investigated this issue a bit and it might be related to uwsgi:

• Django/uwsgi - timeouts after period of time unbit/uwsgi#735
• https://github.com/unbit/uwsgi/issues?q=django%20killing -> issues with Django and killing workers
• https://docs.djangoproject.com/en/5.1/howto/deployment/wsgi/uwsgi/ -> could be worth a look
• https://stackoverflow.com/questions/25414902/how-to-find-out-why-uwsgi-kill-workers -> Issue might also be related with running out of RAM or some process being out-of-memory

Vielleicht ist irgendwo in Django, DOcker, uWSGI Memory auf 2GB begrenzt und das wird immer mal wieder erreicht und dann startet sich die API neu und schießt vorher die Worker ab und das bedingt, dass die Datenbank auch neustartet.