Taking a look at the cluster capacity graphs in the UI, we are not calculating the unused capacity & unused request correctly. 

For example, look at the memory graph in the attached screenshot, our usage is higher than our request, so the unused request should be Zero.

Then taking a look at our request graph. The request is higher than our usage, therefore we should be calculating the unused capacity with the request instead of the usage. The backend work I did to figure out the unused request & capacity for the node capacity epic can be applied here so that the numbers can come out directly in the UI. However, with the cluster capacity we will not be able to get the same counts as in ("8 cores") as we do when group by node, because our UI summary tables don't currently have that information for the cluster level.