Perhaps a better example would be automating a previously manual failover, i.e. if a service becomes unavailable on one machine, start it on the warm standby. O

Perhaps a better example would be automating a previously manual failover, i.e. if a service becomes unavailable on one machine, start it on the warm standby. Or in EC2, terminate an instance and start a new one.