What would you like to be added:
Support kubectl-kruise top cloneset [cloneset name] so that we can see the resource usage of each cloneset.

Why is this needed:
We often want to have an idea of the resource usage of each cloneset, not only individual pods or nodes.

The metrics server might not allow a direct query of this info, but kubectl performing a local summation of the resource usage of all pods in a cloneset may be worthy.