Similarly, impossibility results in distributed systems theory can act as a pretty good cover. Something along the lines of "You know, in a distributed setting you cannot distinguish node failure from delays in communication. So no, I won't be writing monitoring scripts for this service. They will be useless."