I'm very much seeing this at work. We have a few dev-teams who are very responsive and cooperative. You see postgres complaining about queries from their appli

I'm very much seeing this at work. We have a few dev-teams who are very responsive and cooperative. You see postgres complaining about queries from their application, you inform them, they realize the danger behind it and fix it. For these teams, we're perfectly fine providing prod analytics, adjusting parameters or sizing for a bit. Shitty queries happen, we compensate with some compute resources, you fix, we rescale. Entirely great, dev can move fast, we can be lean. And then we have other teams. Some query executed twice for every user transaction has a high chance of resulting in unique key violations, which are logged as important errors. Fixes would be not too hard and have been communicated, but deemed "Not customer visible" and thus prioritized to death. I'll let it up to guesses by the readers which project had sev-0 issues caught early by us and fixed by dev with almost no impact to customers, and which project ended up with a .3% increase in query error rates, which ended up as a sev-0 issue, with SLA discussions involved, and we were entirely blamed for being unable to do shit or at least detect it early. I dislike growing into the operations team I don't want to be, but some teams demand doing so very, very hard.