Hi came here to say , I have sporadically seen similar issues with these devices and other sonoff devices,.
After it all came to a head recently, the fix was to remove a dongle-P (flashed as a router with latest router firmware) from the network. They havent skipped a beat since.
My co-ordinator is a dongle-e. I use ZHA.
Reading some posts above I wonder if there is a zstack issue or a TI related issue but the dongle-P is in a drawer, maybe it will return as a Thread device some day !
ps. I had an automation alert if a thermometer gave me the exact same reading for 6 hours or more so I knew to go and re-pair the thing. I think this is no longer needed now I found the culprit !
pps. actually when I removed that router, I had 3 stuck thermostats. I only re paired 2 of them, the other just started working I noticed. I was just in the zone so used to re-pairing them. My network now seems to self heal, it didnt before.
I know I’m repeating what @buz said here, but it is always a good idea to press the button on battery-powered zigbee devices every couple of seconds while pairing.
This prevents the device from going back to sleep in the middle of the pairing process and avoids issues down the line.
sorry to leave you alone…
the z2m webui works on port 8080 by default, so just direct your browser to <your home assistant IP address>:8080
I have HA and Z2M running in containers, so it’s pretty straightforward for me. I’m not sure it works with every z2m installation.
This was exactly what had happened to me; I’d tried all the other stuff (USB extension, re-flashing the dongle, even trying to use Zigbee2MQTT) and this simple fix worked first time! Thank you so much!