How do you configure and tune Kubernetes Horizontal Pod Autoscaler for a latency-sensitive API service? What metrics do you use and what are the common tuning challenges?

Sample answer preview

The Horizontal Pod Autoscaler in Kubernetes adjusts the number of pod replicas based on observed metrics. For a latency-sensitive API service, default CPU-based autoscaling is often insufficient, and you need to configure custom metrics and carefully tune the scaling behavior to…