You need autoscaling for a stateless, distributed service on GKE with minimal changes and a two-week deadline. What should you deploy?

For a stateless, distributed service on GKE with a tight deadline, a Horizontal Pod Autoscaler (HPA) scaling based on CPU load is the most appropriate and quickest solution. HPA automatically adjusts the number of pod replicas based on observed CPU utilization (or other metrics), which is ideal for stateless services that can easily scale out by adding more instances. CPU utilization is a readily available and common metric for autoscaling, requiring minimal configuration changes. Vertical Pod Autoscalers (VPA) adjust resource requests and limits for individual pods, which is not suitable for scaling out a distributed service by adding more instances. While HPA can use custom metrics, configuring them takes more time and effort, which is not ideal given the two-week deadline.