It could be that it did happen before, but it was just individual cases and if the outage wasn’t long then probably wasn’t noteworthy. This time it was a whole bunch of people affected all at once for a prolonged period. And you’re likely right that there’s probably a series of states the device can be in, and it does calls to AWS as it moves through them, so probably got stuck at a particular stage and couldn’t move forward cause it couldn’t talk to the mothership.