This was the largest AWS outage in a long long time and was still constrained to a single AWS region.

Which is my point.

The same fault on Azure would be a global (all-regions) fault.