I’m running a C8, stable for about a year, multiple Zigbee Inovelli Blue light switches and Z-wave generic blinds with a handful of other small things. I haven’t added anything recently, and have never had the “excessive load errors” previously. No Z wave ghosts- all cleared previously.
I noticed yesterday the smart switch responses getting sluggish, and not working at all today. When I checked the logs I’m getting multiple excessive load warnings from every single zigbee and z-wave device I select (picked enough to see the pattern). I assume this is a result of something overloading the hub to the point of being overwhelmed by any request at all. A reboot seems to have everything working again and watching the live logs for a while shows pretty quiet activity.
Is this something I should look into, or only if it becomes a pattern? I’ve seen lots of posts about checking individual devices for over polling and power reporting (my switches have power reporting turned off), but I couldn’t find mention of this behaviour. Anything else I should check as a possible culprit?
Some process probably got all bound up and the hub could not recover.
I have someone who had this happen last week, hub had been up for 80 days without reboot.
I rebooted, updated and then set up a weekly reboot to hopefully prevent it going forward.
I personally think it is an issue that you will only encounter after a long uptime. I would not worry about it unless it happens a second time shortly after the first.
@gopher.ny@bobbyD why do we have all these log entries but yet there is no alert generated on the hub that there is a problem? I checked the alerts and all it had was an update alert. I have seen at least one other person (other than @DocNaes) have this same issue recently on the forums.
As much as I hate having to do it, I setup weekly reboots for this person to avoid this going forward...
Now, I finally got into the Grafana cloud I setup for this hub.
There does not appear to be any excessive load up until the hub stopped processing events and reporting things up to influx. Unfortunately I did not think to check the actual endpoint for the CPU stats before I rebooted, so what happened between 1/2 and 1/10 is lost.
1/2/25 is when everything stopped working correctly.
That log list looks vary similar to what I found on my hub.
Not home to check the firmware version, but it was up to date about 1.5 months ago.
I may try to auto reboot thing every week or two. I mean it probably won't hurt. I have a home assistant that tends to go offline without a reboot more often than the HE anyway.
I've seen this issue on 2 hubs running 2.3.9 firmware. I came to the same conclusion as you did: it happens after a very long uptime, and I set up weekly reboots.
Probably, how long had it been since last reboot? If you can get into the UI the uptime is on the Logs > Device stats page.
The solution is to reboot or pull power if needed.
The fix for now would be to schedule weekly or monthly hub reboots if you leave it for long periods of time without doing updates or otherwise rebooting.
Oh it definitely had been awhile, and after finding this post I rebooted and its working. My point more is it seems odd that a device needs to be regularly rebooted to function normally. Yeah its an easy fix, but if a person (ie. definitely me) doesn't know these sort of things it could be problematic. But the community here is awesome!
One of my hubs randomly started acting up and high CPU. Reboot didn’t fix it for long. I then rebooted choosing the advance option to do a database rebuild and that solved the problem. Something got corrupted in the database I guess. Might try that too.
IMO it IS expected right now. There have been multiple cases where 80+ days uptime and the hub starts going haywire. One of the people I do remote support for I just set theirs up to reboot weekly so I didn't have to worry about it again.
With the frequency of platform updates most people don't leave the hub running for that long without a reboot to begin with, so you wont see it happening en mass.
I hear what you’re saying but even in these cases there probably is a (rare) underlying cause(s) that, if identified, could fix the need to work around the issue by rebooting?
Yes probably, what I would consider a bug in the platform.
The hub I encountered it on has very little automations, all very basic stuff in RL and Basic Rules, a few in RM. I think 0 IP based devices, all Z devices. I do have Influx logger going to the cloud on it, so possibly that caused some sort of downward spiral. Otherwise very vanilla hub.
