There's normal levels of overengineering, and then there's critical-systems-software-people-making-hardware levels of overengineering. We don't know what we're doing, we know we don't know what we're doing, and we never want to ever have to do this again, but by the gods, we do know the power of redundancy, so we'll just build the entire thing out of redundancies (and then some) to absolutely ensure it can't fail. If it still somehow manages to come down despite all that, it was the gods' will and nothing could have saved it. That or we still managed to underestimate our own incompetence, but that'd hardly be our fault either; someone let software people do hardware, what did they think was going to happen?