The shift from "ops team owns production" to "platform team accelerates developers" is real — but it also changed what monitoring means.

In 2018 you monitored for ops. In 2026 you monitor for developers. The signals need to be different: not just "is it up" but "is it returning the right data." A 200 OK that returns garbage is still an incident.

That gap between uptime and correctness is something most teams still don't close well.