Lg SmartThinQ - washer, start throws error code on switch.turn_on

We have smart washer that is visble in HA.

If I enable the Remote Start from the washer, and a few hours later want to start it from an automation it fails with the error code (switch.turn_on, as the machine has gone into sleep mode)
Error: {‘data’: None, ‘voiceData’: None}

I have tried to mitigate this by adding continue_on_error: true but the automation still aborts.

The relevant action part of the automation

actions:
  - alias: Turn on TM Power if gone to sleep
    if:
      - condition: state
        entity_id: switch.lg_tm_power
        state:
          - 'off'
          - unavailable
    then:
      - action: switch.turn_on
        continue_on_error: true
        metadata: {}
        data: {}
        target:
          entity_id: switch.lg_tm_power
      - delay:
          hours: 0
          minutes: 1
          seconds: 0
          milliseconds: 0
  - action: button.press
    metadata: {}
    data: {}
    target:
      entity_id: button.lg_tm_remote_start

If I manually operate the washer from the device controls in HA it works as intendes, if I try it out in Developer Tools > Actions the switch.turn_on throws an error but the washer is turned on never the less.
I have tried with the integratio wake up, but that fails all together (smartthinq_sensors.wake_up).

Does the integration go into sleep mode or the washing machine does, and ignores wakeup signals?

The machine goes to sleep, by LG design.

It does wake up, and if I operate it from the device in Ha all works as intended.
If I operate it from actions (either from automation or Developer) the power_on and wake_up both generates an error

Trail and horror…
The action switch.turn_on is sent to the machine, and it goes into waken standby mode, however the
continue_on_error: true
is ignored by the automation, which aborts. Relauncing the automation a (few) minute(s) later results in the intended result, ie the washer to actually start.

So, how is the continue_on_error intended to work here? Is it the usage of IF-condition that breaks it?

If I can’t get the continue_on_error to work properly, then the only remaining option is to fire one additional automation based on the exit code of the first automation.

Looks like you found a bug with the integration. Is there a bug listed in GitHub for that add-on? If not report one so the owner can remedy it.

Yeah, I should do that.

As an interim solution I selected to put the switch.turn_on in an one liner script, it does fail, but as it is external from the automation which then survives the error and continues as intended.