Regular 'checkin' events from IKEA action button (even if not pressed)

I have an IKEA action button connected via a ConBee2 USB transceiver.

While debugging some automation I found that the button is regularly sending “checkin” commands.

The events I am listening to look like this:

event=zha_event data={'device_ieee': '0c:43:14:ff:fe:ae:e5:f0', 'unique_id': '0c:43:14:ff:fe:ae:e5:f0:1:0x0020', 'device_id': '1
c025acc2434aea6bdc830a1742ae48d', 'endpoint_id': 1, 'cluster_id': 32, 'command': 'checkin', 'args': [], 'metadata': {'origin': 'LOCAL', 'time_fired': '2021-11-30T21:31:29.219228+00:00', 'co
ntext': {'id': '8abd87d15e9ba63e9914b57c4ed253cb', 'parent_id': None, 'user_id': None}}}

Other buttons of the same type are `silent’.

I would have expected such a device (that does not act as a hub) to stay inactive until one presses a button… What is this ‘checkin’ command? Is it the device trying to find a node in the network? If yes, why is this message showing up on the event bus? That would seem like something that is internal to the network and of no relevance to the application layer…

Im seeing the same, not sure if its recently shown up or something I’ve only just noticed. I caught it by seeing the “ZHA - IKEA five button remote for lights” blueprint/automation restarting over and over in the logs

e.g.

  platform: event
  event:
    event_type: zha_event
    data:
      device_ieee: 84:fd:27:ff:fe:19:d7:e9
      unique_id: 84:fd:27:ff:fe:19:d7:e9:1:0x0020
      device_id: ef81c7e028452c7ae4425d5660237e5b
      endpoint_id: 1
      cluster_id: 32
      command: checkin
      args: []
    origin: LOCAL
    time_fired: '2022-03-02T18:53:36.200041+00:00'
2022-03-02 20:46:55 INFO (MainThread) [homeassistant.components.automation.zha_ikea_five_button_remote_for_lights] ZHA - IKEA five button remote for lights: Restarting

Did you ever find out what this was @encryptix or @has2 ?

I believe the event is caused by the “checkin.interval” attribute of the Zigbee controller, but I have no idea why that particular event would triggering an automation :thinking:

Still seeing it if I check in the history of the Automation but I no longer have warning messages appearing in the logs saying Restarting. I would guess the events lately have been timed further apart to not not queue?