Kubernetes autoscaling is a powerful tool, but many setups fail silently due to misleading metrics, poor configurations, and other details. Here are 10 common pitfalls that can break your autoscaling—and how to avoid them.