Alas, either due to our lacklustre training, or a fat-fingered operative, an unfortunate sysadmin managed to issue a simultaneous, and instantaneous, shutdown to every data centre in every location around the world,
This can't happen with everything in the cloud now can it?
Of course it can - in fact it is a LOT easier - assuming you are using best practice DevOps approaches and have a useless admin with global privileges.
There are plenty of examples of people not just shutting down cloud services accidentally but actually totally deleting them! Get the architecture wrong and it would be easy to do that across every service and location.