Homey Pro 2023 behaves erratically after power outage

My Homey Pro 23 is driving me insane.

Last Sunday, we had a fuse that blew, and the power in the entire house went out.

We isolated it to the fuse group at one end of the house, where we have bedrooms, a bathroom, and a shed connected (unfortunately, also the part of the house where Homey is located).

After a few hours of troubleshooting, we gave up and had an electrician come on Monday to fix the problem.

When we gave up on Sunday, I connected an extension cord to Homey so we could restore automation in the house.

All lights are Zigbee based and are connected to Homey in order to act as Zigbee routers for our door and window sensors to connect to.

This has, of course, caused some interruptions to Homey, hubs, and the network.

All light bulbs in the bedrooms and attached hallway were off until monday - so not acting as routers for our window an door sensors in that period.

BUT since then, our lights (especially in the kitchen and in our bedroom) have behaved VERY strangely - when three bulbs in a lamp group should turn on, maybe only 2 light up… and sometimes the last one turns on after a delay of some minutes.

I have tried to perform a software restart of Homey without any change in behavior.

Our door and window sensors all seem to work fine and fast still.

(How do I see if my sensors have jumped and are now connected to Homey instead of the router/light bulb, they should be using?)

At the same time, Homey suddenly seems very sluggish, and when I open Homey in the desktop browser, there is often a message that Homey is offline/unavailable.

If If I can connect to Homey and I click into flows, flow cards often appear as empty gray placeholders.

What the heck just happened? And equally important… How do I fix it?

Or, as a start, how do I troubleshoot it?

Your issues seem to point to Homey either being very busy with something (which you can check using Insights, System, CPU Load/Clock; those might only be available when you turn on the Power User experiment), or possibly an issue with the USB cable between charger (or Ethernet adapter if you have it) and Homey.

I have ensured the USB cable between the ethernet adapter and Homey is plugged right in and also the ethernet cable.

I think it is fairly evident from the attached screenshot that the outage was on the 19th.

Up until the 19th everything was fine. However, I don’t see anything that seems to indicate an issue after the 19th - but it is obvious that things changed.

A fellow danish Homey user has suggested me to restore from a backup from the day before.

I would be willing to do it, if it fixes the issues but having to do a repair on all my zigbee devices is a bit of a mouthful to put it mildly.

I realize I don’t have the answer as you’ve tried what resolved my issues but I had similar behavior last Monday when we experienced a power outage in our block for an hour or two.

After power was restored everything in homey was super sluggish (lights responding slow, late motion triggers etc). For me it got resolved after doing a graceful restart. Still one or two sensors seem to jump around, trying to get their mesh back in order, but other than that it’s back to normal…

1 Like

That’s not what I meant. The USB cable that comes with Homey is known to cause problems.

I don’t think so. Looking at the top chart something started causing a lot of CPU load on Nov 15, and after the power outage its average is higher than before.

Why would you need to do a repair after restoring a backup?

Yesterday, the only thing I did was to make sure the USB-cable to Homey and the ethernet cable was firmly plugged in.

Then I wrote you here.

Since then everything seems to have been running smoothly (at least we haven’t experienced any hiccups yet).

So I am crossing my fingers and assuming everything is fine now… Though, not really knowing what the cause of the erratic behavior truly was.

I don’t think so. Looking at the top chart something started causing a lot of CPU load on Nov 15, and after the power outage its average is higher than before.

I agree, that on the 15th something spiked the CPU - is there any way to learn what that was in retrospect?

Why would you need to do a repair after restoring a backup?

I just read that Athom wrote this: Some Zigbee devices might need to be send a repair command to start working on your restored Homey Pro. To do so you can go to the device settings on the Homey Mobile App and click “Try to repair” under Maintenance.

You’d have to look through the CPU usage charts of all your apps to see if there was one that might have caused the problem. But if the reason for the CPU usage was system-related, there’s no way to find out what caused it.

Some Zigbee devices, but from what I know (not personal experience) it usually works fine. However, nobody will be able to guarantee that restoring a backup will fix any issues, especially with cloud backups which are very minimal (they just contain data files and configuration, but firmware and apps will always be restored to the current version, not the version you were running at the time the backup was made).

1 Like

Awesome, thanks.

I will look into the data to see if there is anything, I can spot.

This just made me realize that I need to make a flow with Sysinternals, to monitor stuff like “WHEN CPU usage exceeds x percentage” or “WHEN CPU usage on average has Exceeded X percent for Y amount of time”.

This could give me some early warnings of possible issues.

I will leave this open for another day and if everything keeps running smoothly, I will close it.

Thanks for the assistance.