But things can cause widespread outages, like DNS stuffups, or backend updating stuffups, or changing their internal credentials stuffups and a number of other reasons. Just because they have multiple datacentres there are still central points of failure. The site itself worked, just no backend stuff worked.