Any global service has pockets of outage sooner or later. I hate Musk and Twitter and would love to gloat over a dumpster fire, but the truth is that if its fully distributed, runs in a DC, has anycast IP or some kind of CDN front, then its normal to have pockets of users disconnected.

Sharding models can mean the users can be widely distributed: I've had service outage for my google experiences and the guy next to me nothing: its not always about where you are, but who you are, how they shard things.

So whilst I hope this is "serves them right" I also think "normal"