JMP instance not shown in remote-control list after a while, requires restart

[kelnos]

Describe the bug

After some amount of time sitting idle (say perhaps 8-10 hours), JMP stops showing up in the remote-control list of other Jellyfin apps (such as the web client or Android app). If I restart JMP and wait a few seconds, it starts showing up again, until another 8-10 hours have passed.

To Reproduce
Steps to reproduce the behavior:

1. Start JMP on one machine and connect it to a server
2. Try to remote-control it from another Jellyfin app (should work).
3. Wait 8-10 hours.
4. Try to remote-control it again from another Jellyfin app (instance will not appear in the remote-control list at all).

Expected behavior

I expect that the JMP instance will be available for remote control from other apps at all times, as long as it remains on the network.

Screenshots
n/a

Desktop (please complete the following information):

• OS: Debian Linux (bookworm)
• Version: JMP 1.11.1
• Installation Method: Debian package downloaded from the GitHub releases from this repo

Additional context

JMP is running on the same machine as the Jellyfin server, which is always-on and connected to a TV via HDMI.

Log file contains repeated:

2024-11-15 06:24:03.936 [info] unknown @ 0 - JS: Sending web socket message: KeepAlive
2024-11-15 06:24:03.938 [info] unknown @ 0 - JS: Received KeepAlive from server.

But then eventually it ends with:

2024-11-15 06:25:03.934 [info] unknown @ 0 - JS: Sending web socket message: KeepAlive
2024-11-15 06:25:03.935 [info] unknown @ 0 - JS: Received KeepAlive from server.
2024-11-15 06:25:23.049 [info] unknown @ 0 - JS: web socket closed
2024-11-15 06:25:23.050 [info] unknown @ 0 - JS: Clearing KeepAlive for [object WebSocket]
2024-11-15 06:25:23.050 [info] unknown @ 0 - JS: nulling out web socket

[kelnos]

Ah, I realized what is going on, I think: I have a nightly cron job that stops the Jellyfin server, takes a backup of /var/lib/jellyfin/, and then restarts it. It seems that JMP doesn't try to reconnect to the server after that happens. While I am more or less causing this myself, this still feels like a bug with JMP.