What is a common mistake with Kubernetes probes?

Using the same expensive endpoint for all probes or setting aggressive thresholds can cause restart loops and cascading failures.