Open peppy opened 2 years ago
CLOCKWORK_STORAGE_EXPIRATION
can be set in .env
to set the expiry of logs in minutes; the default is 1 week
Hmmm, that seems super weird since I don't think I was actively using my dev environment enough to cause this to happen in just a week. But maybe I left a single chat open polling?
Could probably also change the default to 1 hour or something more sensible.
In the fail scenario, after a few requests, all further requests would block, taking minutes to run (seemingly never timing out until the client closed connection). This would hit the fpm limit and stop all requests from being processed.
I believe the following error may explain the issue, but I'm also not 100% sure that this is the root cause and not a side effect of me restarting the docker deployment a few times rapidly.
Regardless of above, deleting
storage/clockwork.sqlite
and the associated lock files fixes this. We might want to delete these on (docker) startup to avoid these kinds of issues?