I don’t really know what to look for in the logs, but with some guidance maybe we can figure this out. Here’s how it typically goes:
Set up daily recurring auto-post at sub.wetshaving.social. It works for about 1-2 weeks, then stops.
I set the entire server to restart every 3 or 4 days, which makes Lemmy Schedule stable for longer, but still every week or two it will stop working.
The solution is some sort of combination of restarting the Docker stack (Lemmy Schedule and Redis) and/or opening the Lemmy Schedule web interface and logging in, logging out, refreshing, logging back in, whatever it takes. The need to load the web page is odd, like it reminds Lemmy Schedule somehow that “oh yeah, there are posts to post”.
A lot of times when I log in to fix it, the scheduled post does not appear in the to-do list. We have a weekly post that stays in the to-do list, but not until I reload, log out, etc. will the daily post eventually find itself and reappear in the list.
When it starts working again, it typically spits out the posts that it missed, sometimes in duplicate.
Any ideas?
Hmm, that sounds like the component that does the scheduling stops, I’ll try adding some health checks.