HADashboard Mode widget stopped triggering scripts

Since last upgrade (but not sure which part was it: the HA upgrade to 116.2, python upgrade to 3.8.6 or appdaemon upgrade to 4.0.5) the mode widget stopped triggering the configured scripts. I have not been changing neither the dashboard configuration nor scripts nor the entities that are used in those scripts. Those scripts still work ok when triggered from Home Assistant automations or manually from config menu. Below is example of my dashboard configuration:

  widget_type: mode
  mode: standby
  title: "Off"
  icon_on: mdi-power
  icon_off: mdi-power
  icon_style_active: "color: white"
  script: script.heating_standby
  entity: input_select.vicare_mode
  title_style: "font-size: 150%;"
  widget_type: mode
  mode: dhw
  title: Woda
  icon_on: mdi-water-pump
  icon_off: mdi-water-pump
#  icon_style_active: "color: blue"
  script: script.heating_dhw
  entity: input_select.vicare_mode
  title_style: "font-size: 150%;"
  widget_type: mode
  mode: dhwAndHeating
  title: Ogrzewanie
  icon_on: mdi-calendar-clock
  icon_off: mdi-calendar-clock
  icon_style_active: "color: orange"
  script: script.heating_auto
  entity: input_select.vicare_mode
  title_style: "font-size: 130%;"

Any ideas what can go wrong? Nothing shows up in HA log. Is there an appdaemon log I could look at?
I have found an appdaemon service log and it shows the following entry when I try to run those scripts from mode widget:

2020-10-16 15:13:55.794443 WARNING AppDaemon: Unknown domain (default/script) in call_service from stream

What could that be?

Interestingly, after restarting appdaemon it is working again. Hope it won’t happen too often…

I had similar problem:
WARNING AppDaemon: Unknown domain (default/script) in call_service from MHD_automation
But it was directly from appdaemon 4.0.5 automation script.
It happens quite often when HomeAssistent is restarted (appdaemon just right after HA)

After restart of appdaemon again, the issue disappear.

Yes, I’ve just noticed that it is happening after (almost) every HA restart. Have to think about some auto restart of appdaemon or a delayed start of the appdaemon service… But anyway this seems like a bug that should be solved.