Make sure you use a docker image that tracks the stable version of Jellyfin. The official image jellyfin/jellyfin tracks unstable. Not all plugins work with unstable and switching to stable later is difficult. This trips lots of people and locks them into unstable because by the time they figure it out they’ve customized their collection a lot.
The linuxserver/jellyfin image carries stable versions but you have to go into the “Tags” tab and filter for 10. to find them (10.8.13 pushed 16 days ago is the latest right now).
To use that version you say “image: linuxserver/jellyfin:10.8.13” in your docker compose instead of “linuxserver/jellyfin:latest”.
This approach has the added benefit of letting you control when you want to update Jellyfin, as opposed to :latest which will get updated whenever the container (re)starts if there’s a newer image available.
While upgrading your images constantly sounds good in theory, eventually you will see that sometimes the new versions will break (especially if they’re tracking unstable versions). When that happens you will want to go back to a known good version.
What I do is go look for tags every once in a while and if there’s a newer version I comment-out the previous “image:” line and add one with the new version, then destroy and recreate the container (the data will survive because you configure it to live on a mounted volume, not inside the container), then recreate with the new version. If there’s any problem I can destroy it, switch back to the old version, and raise it again.