You can always calculate the theoretical maximum of traffic that can flow through a pNIC.
If you then look at the ratio of the traffic over the VM's vNIC over the pNIC capacity, you would end up with a percentage.
But it becomes more complex when you start teaming (portgroup and inside the guest OS).
And even more complex when you have multiple portgroups running on the vSwitch that has the pNICs mapped.
In a similar fashion you can calculate an IO percentage.
There it would depend on what kind of storage you are using (FC, iSCSI, NFS).