It’s wonderful to be loved, of course, but being the only guy who can save the world while you’re on vacation is a major bummer (see Clark Kent at Niagara Falls). The problem here is once again that what is good in the short term is not necessarily good in the long term – I might save a bunch of money *right now* by only paying *one* guy to know everything there is to know about our enterprise production messaging infrastructure, but when he gets hit by a truck, leaves for greener pastures, or goes on vacation, the world can end in such an amazingly expensive way that in hindsight, *doubling* or even *tripling* our costs by hiring more people to avoid this single point of failure would have saved us gobs of money.
Now part of this can be ameliorated by *useful* documentation (screencasts, step-by-step walkthroughs, or FAQs and HOWTOs actually *built* to answer questions that came up in real life, as opposed to the kind of Word document that lives on a Sharepoint server for decades without anyone actually reading past page 3), but even with all the documentation in the world, if someone isn’t actually familiar with the system (i.e., they’re an offshore “production support” resource that is supposedly an expert in the generic technology at hand but not with the actual implementation in question), they can take orders of magnitude more time to solve an issue than someone steeped in its day-to-day operation.
So what’s the answer, besides having a magical wish fairy that can simultaneously convince business to spend more money short term as well as produce a strong technical candidate who is available for the job? I’d argue that at the *very* least, for each moment of crisis that you *do* actually face, consider that a warning from the fates and solve *that* long term – if it happened once, it’ll happen again. The inclination will be to pass blame around, of course, but resist! Answer the question, “what do we have to do *now* to avoid this in the future?”