This should not happen. This is an “orphaned pod” situation, where JupyterHub believes a user’s server has stopped when it hasn’t. Since you have kubespawner 4.2, upgrading z2jh and kubespawner may fix it. We have had some issues related to orphaned pods recently.
The key bit of logs to find is when JupyterHub is deciding to stop the server, or deciding that the server has stopped. Do you still have the logs of jupyterhub startup before [D 2024-02-15 12:43:14.183 JupyterHub proxy:880]?
There might be something like username appears to have stopped while the Hub was down. That would be the bug, or a similar one, that has been fixed in the latest releases of kubespawner/z2jh.