Fine Tuning of HighOverallControlPlaneMemory Alert

1. Proposed title of this feature request:

• Fine Tuning of HighOverallControlPlaneMemory Alert

2. What is the nature and description of the request?

• Currently HighOverallControlPlaneMemory Alert is generated when the total memory usage across all Red Hat OpenShift Container Platform 4 - Control-Plane Node(s) exceed the 60% mark for more than 1 hour.
• This is a false alert in the clusters where ballooning is activated. The alert compares the used memory of the nodes with the sum of the assigned memory, but this second parameter seems to be the (variable) assigned memory, not the actual maximum memory assigned by the hypervisor. 
• Could it be possible to use the value of the maximum amount of memory assigned by the hypervisor in the calculations? It may be necessary to use what is documented in https://access.redhat.com/solutions/5562311.

3. Why does the customer need this? (List the business requirements here):

•  These alerts are very frequent and spamming the system. Also these are false alerts as ballooning is activated and enough memory is given.