They mentioned having a backup AWS cluster that would spin up when something happens.