For established businesses, I wonder if you could come up with a useful metric for how much value you add by rejecting a commit and thus avoiding an operational incident.
Mature, high-profit systems warrant much greater scrutiny on incoming changes than businesses/systems which aren't spinning off money yet. You can err in both ways: too slow where you need to be fast, and too fast where you need to be more careful. And AFAICT companies/organizations build processes, habits, and platforms around one mode and have a hard time switching to the other.
Yes. Here, every second of downtime is over $10, if it affects all users.
Last time we had a partial outage that only affected a piece of the base for a few hours, we got a first estimate at over 4 million dollar direct loss of revenues, and some percentage of that as direct guaranteed profit loss.
Mature, high-profit systems warrant much greater scrutiny on incoming changes than businesses/systems which aren't spinning off money yet. You can err in both ways: too slow where you need to be fast, and too fast where you need to be more careful. And AFAICT companies/organizations build processes, habits, and platforms around one mode and have a hard time switching to the other.