There is no general recommendation to reboot at x frequency because that doesn’t accomplish anything when a hub is working as intended.
However if a hub is performing inadequately and rebooting seems to help, there are a few likely culprits, as Bobby mentioned.
The key is to understand what’s slowing your hub down, if possible, and either to fix the offending app/driver, or using tools like hub information driver, figure out how to mitigate the effects of what’s slowing your hub down (e.g. by rebooting when free memory reaches a level you’ve determined correlates with hub slowdowns/freezes).
I still have an "old" C-4 that runs hot at different times, sometimes day-to-day, sometimes weeks apart... I haven't really zeroed in on the root cause(s), but am suspecting it is likely down to my API calls or other Community apps / drivers.
Certainly the C-7's seem to run better than the older C-4's, at least in terms of temperature.
In the end it depends on your setup, rather than a specific over-arching approach.
On my end, I have my C7 with several custom apps and drivers that can run flawlessly for months without reboot. However, the only things that I have problems is when I want to update custom drivers or app codes. I'm simply not able to save the update. For that, I have to reboot the hub, even if it was just running for a day.
My recommendation is to obsess about simplicy and not about the outcome of the complexity. If you keep your home automation logic simple, then the hub can run for many months without a single reboot.
We can ensure that your hub reboots periodically, because we release updates regularly.
But the more complicated your set up is, the harder the hub works and the sooner it needs a reboot.
I truly speak of my own experience. Finding the simplicity in home automation is as hard as troubleshooting the complex logic, but it pays off in the long run, time and again.
We run thousands of Windows PCs in our manufacturing plants around the globe. These run for months and months without reboots. The only time they get rebooted are for patches/security updates. Same with my Windows PCs at home.
Admittedly a different use-case and different hardware, but.... well that will produce a different outcome in most cases... Not a criticism....Just an observation that I expect these are very different use cases and hardware / software setups and so will produce different outcomes... I expect...
@bobbyD anything special about C-3 and C-5 hubs and the potential for memory leaks? I'm running Joe Page's Android Dashboard app and it makes use of the Maker API. Just curious and not having any slowdown issues but I do scheduled restarts twice a week.
C-3/C-4 are missing some features that later models have, so they may require more frequent reboots. However, heavy use of Maker API, may cause the hub to slow down over time across all hubs.
Good point... You probably need another two, one to add to your production setup and a development hub.... And maybe another for dashboard development....
I have suspected this as a possible issue ever since I switched to the newer Homebridge V2 app and deleted the MakerAPI integration. It's been a long time ago now, but other changes I made, like adding a second hub, made it hard to pin the improvements on deleting MakerAPI. Keeping KISS in mind with every automation/integration has made it so I no longer have these issues.
I have a rule to reboot on zigbee going down. Seems to be sufficient . That is the first thing to crash under heavy use or low memory the above rule is too restrictive for me . My hubs regularly run with way lower memory . Ie
I've found any integrations involving LAN are trouble, and 1 too many LAN integrations can cause trouble. I think the "too many" varies based on each particular integration, so 3 can be "too many" or maybe in other cases it takes 5
I have to reboot my C4 about 4 times a day. The database goes from ~40mb at boot to well over 120 mb within 4-6 hours and support couldn't figure it out. "Engineering is looking at it" but haven't heard anything in many months