Do the logs of the add-on say something?
Sounds very similar to my issue I’m trying to debug here: https://github.com/dresden-elektronik/deconz-rest-plugin/issues/3735
(Still no idea what’s actually going on, but maybe you can look there if anything of the things proposed there helps you)
Actually I think I managed to solve the issue. I have a 3 floor house, and on the -1 floor is my HA / Deconz PC (it is a NUC with i5 / SSD no speed issue here). On the ground floor I had 6 Tradfri e14 flame bulbs and 6 Immax e14 RGB bulbs. The Ikea bulbs kept falling off the network and wrecking havoc everywhere (since it is literally the middle of the house). This is a recent behavior BTW, it used to be stable a couple of months ago - but I made a lot of changed to the rest of the setup in the meantime so hard to tell for sure. All Ikea bulbs had the current firmware.
I now removed the Ikea e14 bulbs and everything runs smooth for now, stable and all…
Side note : the Ikea e14 bulbs were running quite hot when on - 70-80deg C I would say…
I am a hardware engineer by trade. Inside most of these LED bulbs and actually most walwarts too you have components that can last for 10 years. LED chips rated at 20000 or 50000 hours. And then they have small electrolytic capacitors that have sub 10000 hour specs (even the best brands). And if they run in hot environments they are rated at 2000-3000 hours. The cheap ones 1000 hours. And as they dry out their equivalent series resistance (ESR) increases and the ripple voltage increases which increases the heat loss. And once the silicon runs warm and supply rails have ripple the controllers start crashing and eventually the whole thing dies.
That may be what you experienced. A dodgy bulb that runs warm and gets unstable and as a router in the mesh it creates chaos.
Thanks for the feedback Kenneth, this is totally possible… The only thing is that those bulbs are a few months old only so even with dodgy components and all I am way under 1000 hours / bulb.
The key thing you said was that some of the bulbs were very warm. A small E14 LED bulb should not get warmer than you can hold around it. Even if the mean time to failure is 1000s of hours poor quality can mean that some components fail immediately. It can also be that you have lamps that do not allow the bulbs to get rid of the heat. Heat kills accellerates component death
In the world of reliabily we talk about bathtub curve.
Hi, yesterday, I upgraded my deconz add-on from 6.5.0 to 6.6.0, and I rollback with my backup to 6.5.0.
I use Tuya TRV HY368, and with 6.6.0 release I have this error:
2020-12-10 06:26:21 ERROR (MainThread) [homeassistant.components.websocket_api.http.connection] [2802015104] /sensors/57/config/mode parameter, mode, not available
Traceback (most recent call last):
File "/usr/src/homeassistant/homeassistant/components/websocket_api/commands.py", line 135, in handle_call_service
await hass.services.async_call(
File "/usr/src/homeassistant/homeassistant/core.py", line 1451, in async_call
task.result()
File "/usr/src/homeassistant/homeassistant/core.py", line 1486, in _execute_service
await handler.job.target(service_call)
File "/usr/src/homeassistant/homeassistant/helpers/entity_component.py", line 204, in handle_service
await self.hass.helpers.service.entity_service_call(
File "/usr/src/homeassistant/homeassistant/helpers/service.py", line 499, in entity_service_call
future.result() # pop exception if have
File "/usr/src/homeassistant/homeassistant/helpers/entity.py", line 664, in async_request_call
await coro
File "/usr/src/homeassistant/homeassistant/helpers/service.py", line 536, in _handle_entity_call
await result
File "/usr/src/homeassistant/homeassistant/components/deconz/climate.py", line 115, in async_set_hvac_mode
await self._device.async_set_config(data)
File "/usr/local/lib/python3.8/site-packages/pydeconz/deconzdevice.py", line 30, in async_set_config
await self.async_set(field, data)
File "/usr/local/lib/python3.8/site-packages/pydeconz/api.py", line 112, in async_set
await self._request("put", field, json=data)
File "/usr/local/lib/python3.8/site-packages/pydeconz/gateway.py", line 106, in request
_raise_on_error(response)
File "/usr/local/lib/python3.8/site-packages/pydeconz/gateway.py", line 220, in _raise_on_error
raise_error(data["error"])
File "/usr/local/lib/python3.8/site-packages/pydeconz/errors.py", line 58, in raise_error
raise cls("{} {}".format(error["address"], error["description"]))
pydeconz.errors.RequestError: /sensors/57/config/mode parameter, mode, not available
It seems that in the new version of phoscon, this TRV haven’t any mode attribute…
Before rollback, I tried to delete and reinclude my thermostat with no result.
I have one other TRV but it is eurotronic manufactured, and no issue with this.
I’m trying to add my deCONZ instance to Home Assistance. The deCONZ instance is running on a separate Windows Server computer. Home Assistance (running on a Pi 4) did not have any problems finding the devices in deCONZ. But after I was done with adding the integration (divided the devices into different areas as the final step), all of them come up as “Unavailable”. And in my log file, I’m getting this error message every minute:
ERROR (MainThread) [pydeconz.websocket] Client connection error
What can be the issue here? Some port not opened in the Windows Server firewall? I’ve tried to delete and reinstall the deCONZ integration, but that didn’t help.
I just tried uninstalling the deCONZ integration, then restart Home Assistant, and then reinstalling the deCONZ integration again. But the devices are still “Unavailable”.
After uninstalling the deCONZ integration the devices were even missing entirely!! (joke)
It has nothing to do with what you changed of areas.
Your Home Assistant cannot connect to your deconz at all. It cannot find it. Did the IP address of your windows based deconz change? Does it run in a VM that gets random IP address when it starts?
Ah, I see my first post was a bit vague here. The devices have always been “Unavailable” - not just after selecting areas. I just tried to indicate that “divided into areas” meant that that was the final step of adding the deCONZ integration.
No, deCONZ is running on a Windows Server computer in the native OS (no VM), and the server is set up with a fixed IP address.
This will be fixed with next release of home assistant, its available in current beta
If they are unavailable even if setup works its probably that it can’t connect to the websocket
Its not an issue with home assistant if it can’t connect to the websocket
Great, thanks for the hint. I needed to open port 443 in my Windows Server firewall. Now it works perfectly. 
Great!
ok, thank you very much.
Does anyone have the “Xiaomi Mi Smart Temperature and Humidity Sensor” and can tell me whether it still works with the current Deconz version? For me, it always loses the connection after a few hours if I have paired it.
Maybe the device itself is not the problem.
I recently changed the zigbee channel due to connection issues of several bulbs mostly in the evening.
That led to the total connection lost of all (3) xiaomi motion sensors, which where stable up to that time.
All other bulbs and motion sensors (hue) had no problem.
So I reconnected the sensors. One of them worked fine.
The others worked for a few hours and then lost connection.
For the following 2 days I always reconnected them. Then, suddenly one of them kept stable until tody.
Now I have only one xiaomi left, which still looses the connection. I still try my best hoping that it will work one time like the others.
I think, xiaomis have a huge problem in a zigbee net when the fall into deep sleep. Then the host might not be able to establish a stable connection. Once it catches the device in the right moment, it seem to be ok.
Sounds strange but maybe patience helps.
By the way: The last motion sensor seem to keep the connection up, meanwhile since 20 hours
I have one in use, and I’ve the same issue - by side of others problems 
It seems that Dconz Zigbee ist not really stable…
That’s sad. I bought the sensor last year and until few weeks (or month) ago it worked perfectly. I thought the sensor is broken because I put it into a freezer for few days. But maybe it was an deconz update.