[BUG] ControllerX + AppDaemon Crash with WebSocket Disconnect in Docker Setup

[fir3drag0n]

🐞 Bug Report: ControllerX + AppDaemon Crash with WebSocket Disconnect in Docker Setup

Environment

• ControllerX Version: v5.0.0
• AppDaemon Version: 5.5
• Python Version: 3.12
• Home Assistant Version: 2025.7.4
• Installation Method: Docker (on Unraid)
• MQTT Integration: zigbee2mqtt

---

🔧 Configuration Summary

plugins:
  HASS:
    type: hass
    ha_url: http://192.168.1.113:8123
    token: !secret ha_token
    cert_verify: false
    delay: 15   # was also tested with delay: 30 and delay: 60
  MQTT:
    type: mqtt
    namespace: mqtt
    client_host: 192.168.1.191
    client_user: admin
    client_password: !secret mqtt_pw

Each apps.yaml entry includes valid light: or cover: fields. No misconfigurations are present.

---

❗ Problem Description

After upgrading to AppDaemon 5.5 and ControllerX 5.0.0 in a Docker environment, all ControllerX apps fail on startup with the following error:

TypeError: argument of type 'NoneType' is not iterable
aiohttp.client_exceptions.ClientConnectionResetError: Cannot write to closing transport

The AppDaemon logs confirm the WebSocket connection to HA is established successfully (auth_ok received), but shortly after, the connection drops and is restarted in a loop:

INFO HASS: Connected to Home Assistant 2025.7.4 with aiohttp websocket
INFO HASS: Authenticated to Home Assistant 2025.7.4
WARNING HASS: Timed out [0:00:10] waiting for request: {'type': 'get_states', 'id': 5}
ClientConnectionResetError: Cannot write to closing transport

Attempts to mitigate:

• Verified HA responds correctly to WebSocket auth (auth_ok confirmed via wscat)
• Confirmed no firewall or port issues
• Confirmed token is valid and works manually
• Used delay: and custom AppDaemon app to wait for HA reconnect
• Tried both IP (192.168.1.113) and Docker service name (homeassistant) as ha_url

---

📎 Logs Provided

• appdaemon.log
• error.log

---

✅ Expected Behavior

ControllerX should either:

• Retry initialization cleanly when HA is not fully ready
• Delay entity fetching until HA connection is stable
• Or AppDaemon should not crash on temporary WebSocket transport errors

---

📌 Additional Notes

This may be related to recent aiohttp versions or how ControllerX v5 handles asynchronous entity resolution during init.

Please advise if this is known or if I can assist with more debug detail.

Thanks!

error (1).log
appdaemon (1).log
error.log
appdaemon.log

[fir3drag0n]

See AppDaemon/appdaemon#2399