I’ve narrowed it down to me compulsively updating things… I’ll click to update the schedule stack, and there won’t be an update for the scheduling stuff, but it’ll update redis. Without fail, updating redis gives problems afterwards.

I changed the stack now (as of yesterday) to use a fixed versio of redis, so when I compulsively look for an update, redis won’t change.

To fix it, I just delete the local folders/databases/etc. (leaving the folder with the scheduled events) and start from scratch. I did that yesterday and a test post worked. This time I’ll add in a step of logging out and recreating the jobs from scratch. There are only two jobs so it’s easy.

As far as logs go, there isn’t anything unusual from 00:05Z. The logs are empty going back further than that… probably some sort of log setting I should fix to make it go further back.