This looks like a regression or at least a change in default behaviour, so you never know, maybe incorrectly a dev used a real-time setting in a non RT context. Maybe just check what the settings are on your side.
Good luck with your testing. If you find something please share it here because I am also very interested in finding out more about this issue.
We are struggling to upgrade from 11 because of issues like this, but they really difficult to diagnose and we have been forced to rollback more than once now, each time something else comes up.
We were able to figure out the issue with this, and the problem was one of our own doing.
We have two Asterisk 13 servers setup at different datacenters, and each of our peers register with both servers. The purpose is for high availability. This setup works well, except because Asterisk stores some user-writable information (like voice messages) on disk, we had to have a shared location for these. So we setup a third server that each of the two Asterisk servers connected to over SSHFS. Then we symlinks /var/spool/asterisk and /var/lib/asterisk to the mountpoint.
It’s /var/lib/asterisk that was the problem. I didn’t realize that Asterisk created a SQLite database and stored it here. So both of the servers’ processes were hitting this same file. Once I moved /var/lib/asterisk back to being local, that resolved it. And this works fine, because user-writable data isn’t stored here.