Outside say redundancy across sites (a thing many offer in one way or another) ... there's not much you can DO.
When I used to work with US bank and financial institutions we used to do "disaster recovery tests" and it effectively meant they shut off a whole site and ran on their backups to see what happened.
When I used to work with US bank and financial institutions we used to do "disaster recovery tests" and it effectively meant they shut off a whole site and ran on their backups to see what happened.