Cos it affects a different class of hardware... not the old servers, but the newer and better ones in production.
Hence, it required a specific set of things to align for this outage to occur -- including the fact that 14% of their production servers were down for maintenance, and the load rebalancing that spilt over to the other servers, which triggered the vulnerability, and that became a domino effect that brought down everything else.
I couldn't expand on that X post much, as we're still bound by the character limit per post. So I kept it short and snappy.