I find most cloud users still have to panic at 3am when there is a cloud outage because usually there are knobs to tweak in their application to mitigate the outage. For example, do a region failover, turn off some feature that depends on dynamodb, or push an emergency release to make the application server handle spurious 403 errors from an Amazon backend.

Since Amazon failure modes are so varied, it's impossible to make the above tweaks fully automated. There'll always be a new way the cloud can misbehave, and often your service can at least maintain partial service by being nimble enough.