We see that the kubevirt_vmi_launcher_memory_overhead_bytes metric is reporting incorrect results.
It doesn't report the metric for each vm and is reporting for some of them multiple times.

As part of this bug we will replace the metric and report the VMI requested memory instead.
This is allow us to calculate the overhead by subtracting the pod requested memory and the VMI requested memory.
 

4.17

1. Check the kubevirt_vmi_launcher_memory_overhead_bytes metric results and compare them to the actual VMs.
2.
3.

Not reporting per VM the memory overhead

For each running VMI the metric should report a separate line with the correct memory requested.