Harlan, I agree with Cerebrum and would also like to know what prevented any communication for so long. Transparency, honesty, and integrity are key for a 24/7 operation. If the engineers were "on it" so quickly, I would think that 5 minutes to put a post of "We know there's a problem, we're working on it" would help alleviate some of the (deservedly) negative comments that arose so quickly. Unfortunately, that's what people will probably remember most, even if the other things are fixed quickly and correctly.
As far as "normal support hours", that should only be applying to routine questions and problems. Again, if you're going to require 24/7 connectivity, then "normal support hours" don't apply unless you believe that we only use the application during those times you provide support. Clearly this isn't the case.
And finally, expecting all users to have 24/7 connectivity is also a fallacy (although it looks like you've realized that and will address it). I'm finding this to be an expectation of too many developers. While it may be true for the developer, they need to address the users, not their ideal situation. Let's take a relevant example. When the hurricanes hit in the Caribbean, Florida, and floods in Texas, a LOT of people lost power and connectivity. One doesn't need a lot of imagination to think about other natural disasters that occur in different parts of the world. But, if that means that I can't access applications with relevant information (doctors, emergency responders, communication specialists, etc) that may be stored in your application, then the application is rather useless when most needed.
So, hopefully we're all learning a lesson and we're trusting Harlan and the team to make good decisions based on the lessons learned from this experience. I look forward to seeing those solutions and continuing to use the tool for a long time in the future.