Keeping up with the curve
The thing I've found is most essential to minimizing this kind of crap is actually something management can do something about, at the "expense" of productivity. Active participation in subject-specific mailing lists, forums, newsgroups, etc is the single best way to have staff prepared to anticipate problems, avoid outages, and minimize downtime when they do happen. The challenge is that the time spent on the mailing list *looks like* time wasted when the outages simply don't happen, and better capacity management means upgrades can be put off for months or years longer than otherwise. It's all invisible money saved, and difficult to account for.
