Does anyone know why my ZHA got more reliable?

Configuration Zigbee
1

A bit of a funny post, but basically I’m trying to track down this surprise positive change.

Sometime in the last month or so, my ZHA Zigbee network got almost 100% reliable. I used to have devices randomly drop or not respond to initial actions until they woke up and reconnected, or miss some sensor events (I.e. a motion sensor would detect motion, but not pick up the end and stop detecting motion). It was very annoying but not in a way that made it totally useless, just happened every so often. Sometimes restarting the Zigbee coordinator (an SMLight 06P7) would help, but it was never 100%.

I had resigned myself to migrating to z2m eventually but hadn’t had the time to actually do it.

Then, suddenly, for reasons I don’t understand, perhaps after a recent HA upgrade (I’m on 2026.3, I generally keep up to date, and it’s in Docker) my Zigbee devices suddenly started acting like they were hard-wired. They’ve been working perfectly, not a single missed event or action or automation in over a week.

The issue is I have no idea why, and I’d love to try to understand more.

The only other possibility is that I took out an Aquara temp/hum sensor that stopped working, maybe around the same time, could that have an impact on the whole network?

Are there useful logs or anything I can look at to see if there was a problem or still is? Mostly when I try to look they seem extremely noisy.

Any thoughts on this mystery? Thanks!

2

I had some devices that were really bad on the wifi network that went in and disrupted the zigbee network, after removing the wifi devices my zigbee also became stable

1 Like
3

Yes, there have been a number of threads over the years where a single device destabilizes a mesh. I’m not sure of the mechanism, but I have had similar experiences with both an end device and a router device.

4

I have a big zigbee network, and one thing I noticed after changing to z2m was that some devices are extremely chatty. In my case powerplugs. Swapping out some and reducing reported resolution on the affected devices has helped a lot.

It seems the devices send a new message whenever a value changes, so instead of reporting 230,02V → 230, 40V → 231,04V My devices now only report 230V → 231V
They stil report every time link quality changes, which can be multiple times per second.

So yeah, removing a device can definetly stabilize a network.