weirdcrap

I'm going to mark this as solved. My only thought is this had something to do with the specific kernel version(s) used in the 7.1.x releases as 7.2 the issue has completely gone away with no changes to the hardware.

Well that sucks. Since this at one point worked and since the <custom> option is still present I'm going to open a bug report on this. Maybe limetech will fix it.

In the meantime I have some Pi 2Bs that I can just plug the USB drive into and setup logging remotely that way. It's irritating I have to setup an entirely separate device when this was literally working for me up until last week when I changed a setting not realizing it was going to break my perfectly functional config permanently.

Ok, I don't really understand what was different between my file and the unraid defaults but I seem to have fixed it.

I created a new unraid flash and booted it on a spare pc I had. I just toggled on the local syslog server so unraid would generate an rsyslog.conf and rsyslog.cfg for me. I made backup copies of my existing files and emptied them out replacing them entirely with the default files from the new flash. Saved them and rolled the rsyslog service.

Now when I setup remote logging it is actually logging to the Pi as instructed.

Oh interesting, it seems like removing the sleeps between the rclone mount commands might have actually made it better? I don't recall why I added them all those years ago, I assume to try and resolve a similar issue.
 
I'm going to do some more testing but it seems I'm on the right track.
 
EDIT: Yeah apparently sleep was my problem. 

Alright so I guess the containers were just kind of borked. I removed them and got a "no such container" error. After reinstalling the containers they both check for updates successfully.
Should have tried it sooner but didn't believe anything was wrong with the containers since they were both working correctly.

I had no problems with the commands above that I can recollect. Just be sure to run FIND before you run FIND & DELETE to make sure you aren't getting anything included you don't want.

All of this seems to have been resolved by putting my own router in. Plugin & docker update checks are 3-4x faster than before. pings to hostname's start faster. I'll keep monitoring for DNS issues but I'm considering this resolved for now.
 
A $100 J4125 box from aliexpress and pfSense CE made a night and day difference over that zyxel.

A final update to this.
 
Unfortunately I had to abandon using UnRAID as the WireGuard server as I couldn't resolve this. This issue coupled with my other issue drove me to invest in putting my own router in front of the remote server.
 
I've now built a site to site WireGuard tunnel between my two routers and everything is working exactly as I would expect it to over the WireGuard tunnel. I'm getting my full speed up to the bandwidth limit I set in the rsync command.
 
So TL;DR is don't expect UnRAID's implementation of WireGuard to be able to move large amounts of data without choking. At least not as of this post, hopefully it can be improved in the future.

Have you modified the plugin with custom rsync flags (under the settings)? the -m flag would prune empty directories but doesn't appear to be part of the default flags.
 
I've never used the gather before, only scatter and I've never seen this behavior.
 
You could turn up the logging verbosity to full and perform a dry run to look and see what rsync's doing without actually affecting your shares or folder structure.