Автоматическое обнаружение поломок API

I've been looking into this for a while because I've had a few conversations with people who've had third-party integrations break in production. At first I thought the main problem was just knowing that something changed. But most teams already have some combination of changelogs, integration tests, Sentry, monitoring, contract tests, etc. The part that seems more annoying is what happens after. You get told that something changed, then someone has to figure out: * do we actually use the thing that changed? * which integration is affected? * where is it used in the code? * is this actually going to break anything? * what do we need to change? And sometimes nothing even throws an error. The API returns 200, the JSON is valid, but some business logic is now wrong. So I'm trying to fi…