One hardware host failed, instances already moving to another. Only a few are affected and should be back online soon.