Experiment with a node pool of 8 vCPU / 16GB machines for the Review Apps cluster

I noticed that the current ratio of requested CPU / Memory is approx. between 7.58/13.32 = 0.56 and 6.75/10.97 = 0.61:

Screen_Shot_2019-10-31_at_14.04.01

So I think it would make sense to experiment with a new node pool of 8 vCPU / 16GB machines to maximize the memory usage of the machines.

I've created a new n1-8cpu-16gb-preemp-cos-optim node pool with these specifications.