Something happened a couple of weeks ago, and now Matter devices keep falling off the damn network.
I have a Homey Pro 2026 (migrated from a 2023) that’s been running swimmingly good since I did the migration. But a couple of weeks ago, something changed, some update or … something. A handful of matter devices (mostly MOES matter over wifi switches) starting falling off the network. Not one here, one there, a whole group of them all at once. And not the same group each time, though there are ones that are always in the group.
The real hassle is that repairing them is re-pairing them all. I have to go to each one, reset it back to pairing mode, and re-add it with the QR code. Thank &deity I kept all the QR codes …
There have been no new matter devices added. The network is stable (Unifi environment), everything is on a dedicated IoT vlan, with good 2.4GHz coverage everywhere. From months of perfectly stable to re-pairing a handful (all at once) matter devices at least once a week if not more.
More frustration … There’s one switch that just doesn’t want to be found. It usually takes several attempts, but this time it’s just not finding it. In the app I tried via Homey and via Android - I always ignore the Android popup since that only offers to join it to Google home or whatnot.
One thing happened I haven’t seen before - a red dropdown appeared for the device/switch saying “Subscription registry destroyed”
In general, when repairing any matter device the Homey→Android process searches for the device, finds it and generates matter creds, adds to Homey, then back to white screen of app and repeats that. It finishes with “Adding to Homey” but that never, ever, completes. I have to X out of it. The device has been repaired, but the process never seems to discover that …
All my tests in Home Assistant have been flawless …
I have disabled ALL automatic updates. I also keep track of all updates so I can trace back to some extent whether a particular update might be causing problems.
Yes, I know. But I read that your problems started WEEKS, so presumably before update v13.1.0 was rolled out. But you keep track of what is happening there, which is good. That was my suggestion.
I’ve had the same problem for about a week. I switched almost everything to Matter over Thread. My Homey 2023 loses completely indiscriminately all possible different devices. A repair is difficult to do because the devices are usually not found.
This has been going on for a few days and it’s really annoying. In the meantime I don’t repair anymore, I just boot homey and all my apps again that sometimes helps.
This morning Homey was holding my 50 devices not reachable. After booting, so at the moment everything works.
Might be asked before, but: do you use zigbee devices as well? If so, how many (approx) Matter-over-Thread and how many zigbee devices are paired with Homey?
Please don’t just hope that the issues will be resolved soon—actively contact Athom Support.
The more users who reach out to Athom, the more data they’ll have to analyze, hopefully.
As far as I can tell, this question hasn’t been asked yet, since the thread starter is having issues with Matter over Wi-Fi devices.
Nevertheless, it would be helpful to hear from Athom regarding the issues with Matter over Thread and Zigbee devices; keyword: single chip
So damn weird - that one switch that was just not re-pairing ? It sat there blinking away in pairing mode with Homey never able to find it. Well, an hour or so later POOF, it was back on the network. I didn’t do anything, it just started communicating again.
W. T. F.
And … this morning a bunch are offline again dammit. Heading over to Athom to submit a ticket or whatever.
No, the Matter page doesn’t show much. 0x18b7b2ea is currently not working and needs repairing, which entails re-pairing with the QR code and all that nonsense. 0xb2fb2d58 is the switch that wouldn’t connect, and magically connected by itself an hour or two later last night.
Another weird thing … The switches seem to have new node IDs now, after I re-paired them last night. But not all - the 0xb2fb2d58 was the problem child, refusing to be found, which decided on its own to rejoin the Matter network after a while.
Old node Current node
0x5e712a6f 0x953e7884 Entrance 4btn
0x9c880971 0xccfb59a2 Fireplace 2btn
0xb2fb2d58 same Kitchen back counter
0x3f100fcb 0xaae8af42 Library 2btn
0x248f4bbc 0x926b1d37 Veranda eave socket
0x8d9273d4 0xa069e227 Basement back door light
When I pick “Manage Fabrics” from the 3-dot menu, I see TWO fabrics. One labeled Homey, and an unlabeled one. As far as I know, Homey is the only Matter controller here. Perhaps the unlabeled one is from the 2023 Homey that I migrated from ?
I have 72 devices with Matter over Thread and 8 devices over zigbee.
1 device brand (1 week) new Vaccum cleaner matter over wifi.
Furthermore, there are 3 bridges, SmartThink, Mi-Smart and Connector (SmartRollos) where I don’t know how I integrated them. They’re probably just mirroring.
Is Athom’s tip really to delete all zigbee devices?
No, it was my tip to temporary disable (not possible for app-less zigbee devices) or even temporary delete zigbee devices, to see if Matter reliability improves.
You’ve got to start somewhere to troubleshoot.
(When you re-pair your zigbee devices afterwards, all flows and such will work again)
If that’s the case, the Multi-PAN Thread/Zigbee hardware most probably is not capable of handling (too many) Thread & zigbee devices next to eachother.
Multi-PAN seemed a nice invention one day, but it appeared to be a quite bad idea and it’s best to avoid using both Thread & zigbee with it.
