Having been on the inside of several critical failures, I'd like to give another perspective. For those working on the problem there is nothing as annoying as someone from management interrupting every 10 minutes to learn your current status. As you are identifying a problem anything breaking concentration slows down identifying the problem. So please, until the problem is identified, don't expect more than an hourly "we don't know."
Once the problem is identified and the solution determined, take the first estimate of time until things return to normal that is given by the techies and double it ... in the pressure something is bound to go wrong. The one I remember best is a 4th of July drunk driver taking out a power pole. We'd been discussing adding backup power.
If things are going wrong in off-hours or weekends-holidays, don't expect the techies to have the company's understanding of time zones or remember what form of communications the users may try. Techies tend to be extremely competent in their area of expertize ... and highly upwardly mobile if they can place that in a broader context.
And when they finally identify the problem, have implemented a solution and are willing to give you estimates as to what functions will be available when, half the users are going to read the notice and say "I don't understand what you're saying" while the other half are saying "but are you using xxx or yyy did you consider zzz so it won't happen again." Know which half you fall within, laugh a bit at yourself and give the people working for at least 28 hours on this a break. Monday morning is just fine for the post mortem and future planning - and by then they should be capable of broaden their perspective a bit. It has been highly focused for many many hours.
