I’ve been battling the problem @danekhollas details above since December. I usually attempt to upgrade multiple times after each failure and usually on the third or fourth attempt the upgrade will finally succeed. Unfortunately this equates to quite a lot of downtime as the site is completely down for ~5 minutes after each failure. I have also rebuilt the container a few times since, the last time being a few weeks ago and that has not solved the problem.
The situation was worse again this evening, as the site never restarted and I had no choice but to give up and reboot, which got discourse back online. Presently, on /admin/upgrade, the docker_manager claims it is “Currently Upgrading…” but otherwise everything is back to normal.
This is a self-hosted, standard docker “by the book” setup as per current install docs.