What can cause heap size to fluctuate wildly (hacksaw graphs) on a new 7.8.0 cluster?

Hi

If you are using default jvm.options, you will be using new g1gc garbage collector on 7.8.x

Might check which gc, you are using
Regards