Level locks disappearing for long periods of time

[bkoblenz]

I have built a hub to access matter devices into a control system. I am using Level Bolt locks as part of this and have multiple different physical field locations as well as stuff at home.

My general model is to poll the devices every 5 minutes to know if they are still alive. I have noticed that ALL of my field locations lose access to locks (subscription expired and no new mdns connects or timeouts in the log) for hours at a time. Then things come back and the communication is fine for many hours at a time. There seems to be no rhyme or reason. The field locations are in more "difficult" meaning more walls and distance between the hub and the matter device (as compared to home) but I would expect shorter and more random communication failures if that was the case. I have never observed a problem at home where I have two locks and one hub and a noisier electrical environment.

However, today, I removed one of the locks (Lock B) from the home for a few hours. When I got back I can see that the other home lock (LOCK A) stopped communicating with my hub about an hour after I left. When I got back home after 3 hours and brought LOCK B back into the picture it too did not connect to the hub (not sure how patient I was). I then rebooted the router and both Lock A and Lock B joined in.

I am confident this is a bug in either the Level firmware, CHIP, or the matter-python-server (most to least likely?) I am wondering if the above scenario triggers anyone who is more knowledgeable than I as to what to try or where to look next.

thanks
-brian

[moose09876]

I have 3 Leviton switches running the matter firmware and I'm having issues with them going offline as well, sounds like a very similar issue. If I navigate in homeassistant to the actual device in settings, it will come back online. If I reconnect it to wireless from the ubiquiti console, it will also cause it to come back online. Sometimes they lose their mind for hours, sometimes days, sometimes they work for a month straight. Very frustrating that I can't track down where the actual issue is occurring.

[bkoblenz]

My experience is that when I power cycle the level lock, it comes back immediately, while power cycling my hub does nothing.

I would be interested to know if powering your leviton switches causes things to recover. Maybe there is some common (slightly broken sw) between the level locks and the leviton switches?

[moose09876]

It seems to me like the issue lies in homeassistant due to the fact that just browsing to it in HA can bring it back online. The Leviton app also shows the devices online and communicating the entire time.

[bkoblenz]

In my case, a "reboot" of home assistant does NOT bring things back, but a reboot of the lock does.

…

On Tue, Jun 3, 2025 at 1:09 PM moose09876 ***@***.***> wrote: *moose09876* left a comment (home-assistant-libs/python-matter-server#1113) <#1113 (comment)> It seems to me like the issue lies in homeassistant due to the fact that just browsing to it in HA can bring it back online. The Leviton app also shows the devices online and communicating the entire time. — Reply to this email directly, view it on GitHub <#1113 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADHHYLMVOR56NPRDXZTKR2L3BX6IXAVCNFSM6AAAAABZ6CJBX6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDSMZXGAZDQNRSGU> . You are receiving this because you authored the thread.Message ID: ***@***.*** com>

[agoode]

I've been having the same issue for a few weeks now. My thread border routers are HomePods. I found a combination of two things was necessary to fix this:

1. Make sure IPv6 with prefix delegation on the LAN is set up. Originally I was using a unifi gateway which doesn't do PD on the LAN. The HomePods would generate additional prefixes and change IPv6 addresses periodically. I don't think this was the biggest issue but made things worse.
2. Most importantly I made sure the Thread network was on channel 25 and my wifi channel 11 was disabled. This seems to have fixed the problem more or less completely. It looks like the Level devices are susceptible to wifi interference. I used this doc to understand it (ZigBee and Thread channels are the same since they both share the same underlying radio protocol, 802.15.4.) https://www.control4.com/docs/product/zigbee/best-practices/english/latest//zigbee-best-practices-rev-e.pdf

Ultimately this was made difficult because of the weak diagnostic support for thread (especially compared to zigbee). The matter server logs were very helpful to see when timeouts were happening but I had to go looking for it instead of getting the issues automatically surfaced through notifications or other means. I think Thread 1.4 may improve this somewhat.

[moose09876]

My issue has been resolved for a week now. Updated Ubiquiti 7 Pro AP's to newer firmware. From ubnt:

UI Support (Ubiquiti Help Center)

I hope you are doing well. Before I flag this to the product team, can you please try updating the UAP to 8.0.43 and checking if the state persists?

We have a few improvements on this build.

Fixed IPv6 leakage between SSIDs in certain scenarios.

[bkoblenz]

I've been having the same issue for a few weeks now. My thread border routers are HomePods. I found a combination of two things was necessary to fix this:

I much appreciate the feedback. I am not sure I fully understand your your comment so I will provide a little more context and would appreciate any further thoughts.

My set has a local network with no LAN involved. In particular, I run my app on a raspberry pi with openthread-border-router running on that same pi. There is a physically attached RCP nrf dongle. Sometimes the pi uses wifi, sometimes ethernet, and sometimes there is a cellular dongle connected to the pi. The main point, is that in this case, I do not have any other any non-local networking involved except the path to the internet.

1. Make sure IPv6 with prefix delegation on the LAN is set up. Originally I was using a unifi gateway which doesn't do PD on the LAN. The HomePods would generate additional prefixes and change IPv6 addresses periodically. I don't think this was the biggest issue but made things worse.

No homepod as referenced above, but might there be a special config on my pi that I am not setting? Recall in my original case, I can reboot the pi and it does nothing, but rebooting the level lock DOES result in it coming back.

3. Most importantly I made sure the Thread network was on channel 25 and my wifi channel 11 was disabled. This seems to have fixed the problem more or less completely. It looks like the Level devices are susceptible to wifi interference. I used this doc to understand it (ZigBee and Thread channels are the same since they both share the same underlying radio protocol, 802.15.4.) https://www.control4.com/docs/product/zigbee/best-practices/english/latest//zigbee-best-practices-rev-e.pdf>

I will check your reference. Thanks. I have this problem also when using a cellular dongle and, in that case, wifi is not operating, though it might be powered on. My setups also tend to be in very quiet electrical environments and there are unlikely to be other wifi signals within 100m.

Ultimately this was made difficult because of the weak diagnostic support for thread (especially compared to zigbee). The matter server logs were very helpful to see when timeouts were happening but I had to go looking for it instead of getting the issues automatically surfaced through notifications or other means. I think Thread 1.4 may improve this somewhat.

What I can see in the otbr logs is the br trying to reach the lock, but no respnse from the lock.

[bkoblenz]

My issue has been resolved for a week now. Updated Ubiquiti 7 Pro AP's to newer firmware. From ubnt:

UI Support (Ubiquiti Help Center)

I hope you are doing well. Before I flag this to the product team, can you please try updating the UAP to 8.0.43 and checking if the state persists?

We have a few improvements on this build.

Fixed IPv6 leakage between SSIDs in certain scenarios.

I dont understand how this ties into my environment. But hopefully it will help someone else.

[agoode]

I spoke too soon and my lock has dropped off again. Thanks to the unifi firmware tip, I've upgraded my firmware and we'll see what happens.

More generally, I think this comment is very helpful:
home-assistant/core#134231 (comment)

There are still some pretty big issues with thread and border routers. The comment above discusses the issues and some workarounds.