This is something which puzzles me as well. Everybody will understand that a constant availability might not be achievable and so an issue once in a while isnt a problem, but outages of this scale are different to comprehend for a site of this size. The only explanation I could come up with is...