A 10-minute lightning talk targeting Crossplane operators and platform engineers on the most common circuit breaker implementation mistakes in distributed reconciliation loops, and the principles for fixing them correctly. Focus on shared state, observable transitions, and time-aware backoff logic.

Most teams implement circuit breakers reactively—trying to "fix" failures—but end up creating worse problems by resetting state too aggressively, storing state in-memory only, or not coordinating across multiple controller instances. The right fix requires: persistent shared state, time-aware transitions, and observable metrics.