Hub lost all zigbee twice in one week

Twice this week my c7 hub lost the zigbee network.... None of my zigbee devices were responding. When I realized it wasn't just one device, I tried stopping and restarting the zigbee service, but that didn't help. Had to reboot the hub and all went back to normal. The second time i just immediately rebooted the hub and that fixed it.

Curious if anyone has an idea... This hub has been problem free for at least 2 years

I am one update behind the current release.

What do the logs have to say? Typically when something is banging the hub too hard, zigbee is the first to shut down to conserve resources.

Look at appstats as well, they will usually have a clue.

Hadn't thought of checking the log... But now that I have, it doesn't seem to show anything unusual other than at a certain time a bunch of devices started reporting "no response".

I also have the device activity check app.. and after a while it did report that several devices seem to be offline. Since what salt it was rebooting the hub, is there any way to programmatically reboot the hub if this condition is detected?

You can reboot with the zigbee radio goes off line by using a hub event trigger but that doesn't solve your issue, it just bandaids it. Start with appstats, disabling all 3rd party integrations and see how things behave. Then start rennabling things., I would also use webcore graphs and Hubitat Information Driver to monitor memory and cpu levels to see if they're dropping to dangerously low levels when the zigbee failure occours. This can be handy in diagnosing the issue. You can also tag one of the support team for assistance and they can look at your engineering logs (you do not have access to them)

As Rick noted, look at Logs>Hub Events - do you see any Zigbee Radio entries there?

The Device Stats and App stats tabs may have some info - look at the usage levels on both tabs:
image

On your Past Logs tab, use the Level option to filter for Errors and Warnings to see if any devices/integrations are running amuck:

image

The log-hub events just shows the two instances of "Zigbee radio is offline" each followed by my reboot

I find it odd since I haven't made any significant changes to devices or network in many months.

I did find that a few devices (Third reality power plugs) were reporting way more power events than needed, so i adjusted the thresholds today.

Also lot a ton of "queue full" errors on some devices but they were all AFTER the zigbee failure, so I guess that makes sense.

That can be an issue, power reporting can overrun the hub a bit. But if it's been the same level of reporting for months that makes it less likely to be an issue. But good idea to adjust to the level you need, rather than run on defaults that usually way over-report.

If you continue to have problems you could temporarily just disable power reporting to rule that out completely.

What is your current hub platform version? Did you update recently?

The only devices that I actually have power reporting turned on are the ones that I'm using power levels for certain rules... Really only three or four devices.

I was One update behind but I just applied the latest update to the hub today.

1 Like

I assume that means you're on 2.3.8.140.

That had some fixes that should help with hub load in general.

Dam Rick, you run your hub at 150K ? Mine slows to a crawl at anything under 250K and the page load time is unbearable.

1 Like

When a C7 was my main hub I also used to be fine down towards 150 - 160mb - I wouldn't know it was that low unless I looked. W/C8-Pro now as my main hub memory is never an issue, and my now lightly loaded C7 never goes below 400. Lots of YMMV going on w/this stuff...we all have different devices, device types, integrations, environments, etc.

In my working life when I worked w/teams trying to address/minimize customer Wi-Fi issues I thought that was a PITA - this stuff can be five levels messier to troubleshoot.

Mine has always been between 140 and 180. I've never noticed any slow downs what so ever. I've had my hub down to 87 before I noticed delays (this was way back when when memory leaks were serious). I have close to 200 devices and maybe 30 or so apps. So far no real issues. Even during the beta.

Just happened again yesterday... Zigbee shut down. Reboot didt solve it, I had to restore (no changes) then all was ok

I took a screenshot of the log just before the first device reported a problem...

Any ideas?

This sounds like something that the support team should look at. @bobbyD