A newly migrated application in an Amazon EKS cluster immediately scales out to its configured maximum number of pods on deployment, even before any user traffic arrives. Other applications in the cluster operate normally. Which change will fix the application's scaling behavior?

The application scales out immediately to its maximum pods, indicating a resource request issue rather than a traffic-based scaling problem. The Vertical Pod Autoscaler (VPA) automatically adjusts resource requests and limits for containers based on their actual usage, preventing over-provisioning or under-provisioning. In this scenario, the application likely has very low or zero resource requests, causing the Horizontal Pod Autoscaler (HPA) to incorrectly perceive high resource utilization and scale out aggressively even without traffic. VPA would correct these requests, allowing HPA to function properly based on actual load. The Horizontal Pod Autoscaler (HPA) scales pods based on metrics like CPU utilization or custom metrics, but it relies on accurate resource requests. If requests are too low, HPA will scale out prematurely. The Cluster Autoscaler adjusts the number of nodes in the cluster, not the pods within a node. The AWS Load Balancer Controller manages AWS Elastic Load Balancers for EKS services and does not directly control pod scaling behavior.