Like I stated earlier, a business in which crisis are so frequent as to occur daily or weekly, is badly mismanaged and dysfunctional. Such companies will hide under false pretenses like "that's just the way the industry works", while frankly straight up abusing developers. Crisis should by all means be extremely infrequent, otherwise there's a major management problem, which of course can't be the developers responsibility. I'll restate from all my replies in this thread, that if OP is in such a company, he should leave.
2) In the variety of European countries where I've worked, compensation for over-time is a non-issue - I did not mention compensation explicitly, as it's obvious to me - it's indeed against the law not to compensate. You'd either get full-pay, 50% to 100% increase (usually for night and weekend work), or you'd get time off. From yours and other responses, I gather this is not the standard in the US.
3) Certainly there were deployment procedures, however this developer decided to bypass staging, testing and even didn't perform a smoke test which would have shown the crashed application. He deemed his code of such high quality that nothing could possibly go wrong. And while he didn't follow procedure and therefore caused a minor scandal for this particular company's public image and lots of lost customers, he didn't see why he should be involved in recovery, since it was after 5 o clock. Instead his teammates worked until long into the night, to find out what had actually happened before finally fixing it.