Question

A platform team runs a stateless web service that experiences unpredictable traffic spikes throughout the day. During peak periods, response times degrade because there are too few replicas, while at night the fixed replica count wastes resources. The team wants the number of running Pods to increase and decrease automatically based on observed CPU utilization. Which cloud native autoscaling approach directly addresses this need?

Accepted Answer

Configure a Horizontal Pod Autoscaler that adjusts the replica count of the Deployment based on a target CPU utilization metric

Answer

Configure a Vertical Pod Autoscaler to raise the CPU and memory requests of each existing Pod during peak periods

Answer

Enable the Cluster Autoscaler so that new worker nodes are added whenever traffic increases

Answer

Manually run a scaling command each morning to increase replicas and another at night to reduce them