ZHA-toolkit in HA and zigbee lights ping system

[Fordy113]

Hello,

I'm encountering repeated issues with the ZHA 🧰 Toolkit custom integration and Zigbee network stability running under Home Assistant (version 2025.10.3, Python 3.13 core).

Below is a summary of all the error messages and warnings from my logs:

---

Main recurring issues

• Device has re-joined the network

  Exception 'Device has re-joined the network' for service call with  {... 'zha_toolkit_version': 'Unknown', 'zigpy_version': '0.83.0', 'zigpy_rf_version': '0.14.1', 'ieee_org': '<entity_id>', ... 'errors': ["DeliveryError('Device has re-joined the network')"], ... 'success': False}
  

  These exceptions happen for multiple Zigbee devices after executing ieee_ping commands via zha_toolkit. This usually logs several times per minute for several entities.

• TimeoutError on zha_toolkit ping

  Error executing service: <ServiceCall zha_toolkit.ieee_ping ...>
  Traceback (most recent call last): ...
  TimeoutError
  

  This indicates some Zigbee lights or switches simply do not reply in time; sometimes closely clustered with "Device has re-joined the network".

• RuntimeWarning: coroutine 'register_services' was never awaited

  /config/custom_components/zha_toolkit/__init__.py:927: RuntimeWarning: coroutine 'register_services' was never awaited register_services(hass)
  

  This warning appears each time zha_toolkit is (re)loaded.

• Blocking call to listdir/read_text/open in event loop by custom integration zha_toolkit

  Detected blocking call to listdir/read_text/open with args (...) inside the event loop by custom integration 'zha_toolkit' at custom_components/zha_toolkit/utils.py, line 42: HA_VERSION = version("homeassistant")
  

  This hints at sync file operations from the toolkit (os.listdir, etc) inside the asyncio main event loop.

• Zigbee hashed link key for unknown device

  Skipping hashed link key ... for unknown device ...
  

  Appears intermittently for various Zigbee device addresses.

• Supervisor failed session validation

  Failed to call /ingress/validate_session - (ERROR)
  Client error on /ingress/validate_session request 401, message='Attempt to decode JSON with unexpected mimetype'
  

  These errors are tied to the Home Assistant Supervisor and web ingress, possibly due to timeout or authentication failures.

• Waiting on integration to complete setup

  Waiting on integrations to complete setup: {('zha', '<integration id>'): <timestamp>}
  

  Seen often on boot, especially when Zigbee network is large or physically unstable.

• Bluetooth fallback warnings

  Successful fall-back to passive scanning mode after active scanning failed (4/4)
  

  Likely unrelated unless using Bluetooth.

---

Impact

• Dozens to hundreds of these error messages can appear in a short time.
• Ping monitoring of Zigbee devices sometimes fails and/or produces lots of delivery timeout errors.
• The toolkit version is often "unknown" in logs.
• Occasionally, ZHA waits a long time to finish initializing and some devices go into rejoin/timeout loops.

---

Request for help

• Has anyone encountered repeated “Device has re-joined the network” and TimeoutError messages from zha_toolkit?
• Any advice for fixing or silencing the blocking file-call errors in custom_components/zha_toolkit/utils.py?
• Is there an update or workaround to improve how zha_toolkit handles join/rejoin and sync operations in the latest Home Assistant core?

Thank you!

---

Источники
[1] Complete Zigbee integration failure https://community.home-assistant.io/t/complete-zigbee-integration-failure/796334
[2] Zigbee Home Automation (ZHA) https://www.home-assistant.io/integrations/zha/
[3] Zigbee Integrations Fails To Set Up Although It Is Working home-assistant/core#99951
[4] My Zigbee integration stopped working and I'm seeing this ... https://www.facebook.com/groups/HomeAssistant/posts/4120276914910314/
[5] Problem adding Zigbee Home Automation integration. https://www.reddit.com/r/homeassistant/comments/12mbpl7/problem_adding_zigbee_home_automation_integration/
[6] Adding zigbee integration stops with error https://community.home-assistant.io/t/adding-zigbee-integration-stops-with-error/371314
[7] Connection problems with ZHA - Zigbee https://community.home-assistant.io/t/connection-problems-with-zha/762314
[8] How to Troubleshoot Zigbee Issues in Home Assistant https://www.smarthomejunkie.net/how-to-troubleshoot-zigbee-issues-in-home-assistant/
[9] Solving Zigbee Problems in Home Assistant | Ultimate Guide https://www.youtube.com/watch?v=S59UlboptaM
[10] Zigbee 3.0 Failed to Set Up · Issue #93766 home-assistant/core#93766
[11] Snimok-ekrana-2025-10-18-v-17.45.50.jpeg https://ppl-ai-file-upload.s3.amazonaws.com/web/direct-files/attachments/images/73931164/f46ea69c-f71b-4b28-94e9-09203c662120/Snimok-ekrana-2025-10-18-v-17.45.50.jpeg

[mdeweerd]

Hi, here is some feedback:

• Device has re-joined the network

My guess is that for some reason the device in question performed a join request during the ping.
Normally devices do not rejoin all the time - maybe they do after a zigbee network restart but I did not check the docks.

https://github.com/search?q=repo%3Azigpy%2Fzigpy%20%22Device%20has%20re-joined%20the%20network%22&type=code

• TimeoutError on zha_toolkit ping

This indicates some Zigbee lights or switches simply do not reply in time; sometimes closely clustered with "Device has re-joined the network".

If you are pinging the lights, I agree. Which makes me think: the lights are always powered? They might also be rejoining because they lose power and then they rejoin when they regain power.

• RuntimeWarning: coroutine 'register_services' was never awaited

Ok, an await was missing, I added it (for next version.

• Blocking call to listdir/read_text/open in event loop by custom integration zha_toolkit

Avoiding this requires a rewrite to read contents asynchronously from a file which always exists.
The home-assistant rules blindly require that there is nothing blocking in the startup. IMHO this is not an issue here as this is fast enough and happens only at startup.

• Zigbee hashed link key for unknown device

Seem to happen when the IEEE address of the "partner" device is missing. I did not check the detail, but devices can have a parent device or a device for their route.

The link to the re-join could be that the device in question did "disappear" from the network.

https://github.com/zigpy/zigpy-znp/blob/fd9f7eb65b7d3262c739d864a7ea483234996999/zigpy_znp/znp/security.py#L300

• Supervisor failed session validation

I do not think this is related to zha-toolkit.

• Waiting on integration to complete setup

This is 'zha' not 'zha-toolkit'.
'zha' depends on hardware initialisation or even the initialisation of an add-on if this is what is used to manage the zigbee network (for instance: "Silicon Labs Multiprotocol").

Impact

• Dozens to hundreds of these error messages can appear in a short time.
• Ping monitoring of Zigbee devices sometimes fails and/or produces lots of delivery timeout errors.
• The toolkit version is often "unknown" in logs.

  I did not see an example of that.

• Occasionally, ZHA waits a long time to finish initializing and some devices go into rejoin/timeout loops.

  Seems to imply that the rejoin/ping issues happen at startup, which can be a reason for devices to "rejoin".
  Pinging devices during startup can help cause the delay.

I guess a workaround is to hold-off the ping after a reboot (wait at least X seconds after a HA restart) - this can be a condition in the automation (require that the ha restart event was more than X seconds ago).