Making a fool of myself over Lemmy - because I feel like the major Instances are not openly publishing their lemmy_server error logs & nginx logs. These "very fast 500 errors" are still ongoing 0.18

Even though 0.18 is installed on Lemmy.ml - the code is failing internally. And without access to lemmy.ml's server logs, I'm trying to diagnose the whole design from a remote instance.

Lemmy.world, Beehaw, Lemmy.ml are all throwing "fast nginix 500" errors on their front door on a regular basis. And all are showing symptoms of replication failures sending messages and content to each other (missing posts and comments).

Even AFTER lemmy.ml was upgraded to 0.18, I was able to get stuck 'pending' subscribes on both my own personal remote instance and over at Lemmy.world:

I've been making a fool of myself as best as I can in hope somebody will step back and actually share their lemmy application error logs on where the faulty points are within the code. I highly suspect that PostgreSQL is timing out or http federation timeouts are happening within the Rust code.