How frequently is needed a reboot?

I would ask to the community how many times it’s mandatory a reboot of the system.
Actually after spent some days to fix and improve the system all worked fine for weeks.
Just until some days ago.
After a reboot on the Home assistant interface all is up again, but I didn’t find any errors on docker console, so the question : how to try the origin of the problem ?
(interface was up but it didn’t control anything)
(I think was the philips_hub : every months (~) needed turn off /on cycle but even that was needed a restart of Home assistant)
Now is up again of course… I don’t know if is there a reboot schedule command, or if is something that occurs and I just live with it.

I never reboot my system “mandatory”, only when needed (updates, changes that require reboots etc). If you have issues start by looking through the logs.

Hard to say. A couple of years or so I used not not reboot for several months and things were mostly fine. I usually rebooted when I changed something in the config or updated HA. With most recent versions I need to reboot more often, mainly due to more and more ‘diffuse’ long-term memory leaks somewhere within HA that I just can’t pinpoint (I suspect the stream component). I’m running on a 1GB Pi3, so the smallest leak will quickly become a problem. I now auto-reboot every month.

Edit: to clarify, with ‘reboot’ I meant HA restart. I don’t reboot the host.

I run my HA in Debian via HA Container (stand alone Docker) on a NUC PC and my reboot sensor says I last rebooted 9 months ago. And I think that’s when I updated from Debian 9 to Debian 11.

I think the time to reboot before then was also in the same ballpark if not longer.

So the answer to your question is almost never.

Now HA restarts are a completely different matter.

And recently even that has begun to be less often. I think the longest time between restarts I remember was ~2 weeks but that was very unusual. Usually its every few days on average just simply due to development.

Yes, definitely I am thinking Restart as reboot but RESTART is the procedure/command I was referring to.

9 months is really pretty good result :smile:
on Docker the “stable” version seems really rock solid, but not enough… unfortunately even logs not are so helpful…