Many items in the kube objects we receive and store can be in slices.  The order of these slices do not matter.  They are essentially sets stored as slices.  We need to ensure we ignore the order of these slices when computing the hash function to avoid flagging a difference for an object that is essentially the same.  

For instance the ordering of subjects within a kubernetes role binding is insignificant.  As such we should ignore that.  This applies to many kubernetes object.

See the below PR where we encountered and resolved this with roles.

https://github.com/stackrox/stackrox/pull/8058