How can you enable horizontal pod autoscaling in OpenShift, and what metrics does it use by default?

• A. OpenShift automatically scales all deployments without configuration.
• B. Create an HPA resource (or HorizontalPodAutoscaler) that scales based on CPU utilization by default, or custom metrics if configured.
• C. Use a Deployment with a dynamic replica formula built-in.
• D. HPA scales based on memory usage only.

Answer: (B)

Horizontal Pod Autoscaling in OpenShift is configured by creating an HorizontalPodAutoscaler that targets a specific deployment (or deployment config) and defines how many pods to scale between. The default behavior is to scale based on CPU utilization, looking at the average CPU usage across all pods relative to each pod’s requested CPU. If the observed usage rises above the target, more pods are added; if it falls below, pods can be reduced, all within the defined minimum and maximum replica bounds. You can also switch to custom metrics by configuring the HPA to use metrics from the Custom Metrics API (for example, Prometheus) to scale on other measurements. This is why the correct approach is to create an HPA resource that scales based on CPU by default, with the option to use custom metrics if configured.