A well run service run by a world-class team will provide global availability performance and service that small teams are incapable of.

However, in this case, I don't think the service is very well run, so there's downtime. Not really an issue with "the cloud", this is just an issue because of bad management and/or design of the system.