You must deploy a Kubernetes workload with unknown and variable CPU/memory needs. You need cost-effective recommendations for resource requests/limits and ensure workload availability across conditions. What should you configure?

The Horizontal Pod Autoscaler (HPA) adjusts the number of pod replicas based on observed CPU utilization or other custom metrics, directly addressing workload availability by scaling out (or in) as demand changes. The Vertical Pod Autoscaler (VPA) analyzes historical resource usage and provides recommendations for optimal CPU and memory requests/limits for containers. These recommendations help to right-size pods, improving cost-effectiveness by preventing over-provisioning while ensuring sufficient resources. The VPA can also automatically apply these recommendations, though for "suggestions," it typically operates in recommendation mode. The Cluster Autoscaler scales the underlying cluster nodes, not individual pods or their resource configurations, so it's not directly for pod-level resource suggestions. While it contributes to overall availability, it doesn't manage pod replicas or resource requests/limits.