Scaling on CPU is a trap for event-driven apps. Learn how to use KEDA to scale Kubernetes pods based on Kafka lag and Redis queue depth with real-world configurations.