As a customer of Red Hat's OpenShift-as-a-manged-service offerings (ROSA, ARO and OSD) I expect that the Ingress services autoscale in response to traffic and load.  

OpenShift Custom metric autoscaler (CMA) enables users to horizontally autoscale pod based on custom metric. 

This Jira feature tracks the use of CMA to autoscale ROSA Ingress in response to a custom metric.  

This feature will be executed in phases. 

The goal of the P0 phase is to provide GA and configuration support of CMA to scale (the number of) ingress pods to be equal to number of infra nodes in the ROSA cluster on day 2. Out of scope for P0 is allowing ROSA (external) users to use CMA to autoscale any application pods for production (that will come in a later phase or tracked separately).