Being a professional developer is hard. That's why it's generally a high paying position, because you solve hard problems and are accountable for the results.

So stop making excuses. An organization of this scale should be taking serious precautions (against code errors, bad hardware, malicious attacks, network problems, compiler bugs, cosmic ray induced bit errors, etc.) when building a credit card processor.

If something truly extraordinary comes out to justify this problem, I may be sympathetic. But Occams razor suggests this was avoidable (through good process, not "I should have seen that off by one error").

I work in banking and when a piece of our code went down due to failing to handle load correctly during a market announcement the business was able to say they lost, in that 7 minute period ~£90k.

That is why companies should be made accountable for software errors, just like most industries.

Your code has bugs, my code has bugs, but when our organization fails to implement a system to effectively catch bugs - particularly catastrophic ones - it's a bit dismissive to throw a "shit happens" at it.

Especially when the bug very feasibly might not be in your system.

In this case, Twilio is obviously calling out to a billing provider of some sort. I've run into situations myself where the billing provider returns a status message equivalent to "failure" when it actually was successful. In that case, many billing systems will try again thereby double-billing or worse.