Sorry for a long delay in our response. I'm afraid that due to some limitations in ESX and our MP, in v 5.7 you cannot see individual metric for a VM which using a virtual disk on your NFS datastore. Although you can see which host is causing the most trouble. If your hosts connected to a single NFS datastore at a time, this could be a way to get to VMs.
We are working to significantly improve discoverability of such information in future version of our MP. Stay tuned.