Exactly, I've had cases when half-assed "backup" components led to cascading failures that were horribly difficult to troubleshoot.