Recently, the lights in my office have been becoming randomly unresponsive, with no change to the setup/environment. They’re the only zigbee devices in the mesh that behave this way, and they’re also the closest to the gateway.
HA, Z2M, and NodeRed don’t show any errors, and the states from the lights/Z2M are toggling, but the bulb itself doesn’t do anything. For example, the light is physically on, but HA says it’s off. I have to turn the light on in HA, and then when I turn it off the light also turns off.
The bulbs are all Tradfri (one is an RGBW, one standard ccww, the other is a large globe ccww). They are in an HA-defined group and are also controlled by the Adaptive Lighting add-on (other lights controlled by adaptive lighting do not experience this issue)
HA and Z2M are up to date, no firmware updates available for the bulbs, and the controller is a cc2652-stick from Slaesh. Everything is running in Docker on a Debian server. HA is the Container/Core version, not HA OS. The dongle is on a short extension and sits above the server and networking gear.
I’m at a loss as to why these would suddenly begin misbehaving.
Ninja edit: I realized that I did add an aqara 4-button switch in the room recently, but it’s not used for those lights. But that’s the only change I’ve made to the zigbee network recently. I’m fairly sure it was added after I noticed this issue, though.
https://www.metageek.com/training/resources/zigbee-wifi-coexistence/
Maybe? Tweaking my channels solved all my random drops.
I did channel switching a while ago, and nothing has changed in that area.
There’s a possible bug in the latest Z2M release, which lines up with when my troubles started. Sounds like a fix is in the works, so hopefully that will resolve things.