For the a complete storage breakdown case you should have reliable backup and a fast recovery process. You need to determine upfront how much content changes you're allowed to use in case of such a total breakdown. Because you cannot backup each and every action, but usually you backup only once (or maybe twice) a day. And btw if a single systems breaks there are lighter processes, which allow you to restore them. On authoring you should setup a cluster to allow restore of a single node without loosing content. On publish you normally just would clone a running instance to replace the broken one (backup a running one and restore it on the broken one).
Site slow: Well, if a site performance degrades, there is a reason for it. Either your developers do not care about performance, then you should have some performance analysis and performance tests. You should have good performance tests. And so on. And performance isn't caused by a single action or activity performed well, but a good performance relies on a number of factors. Which aren't relevant for any BCP in the first place, but you development process should take care of it.
Jörg
