$ kubectl autoscale deployment <deployment_name> --min=<min_replicas> --max=<max_replicas> --cpu-percent=<cpu_percentage>
Set up autoscaling for a deployment based on CPU usage.