I've found many times while trying to debug spurious problems with incoming data that it would be very helpful to track the real new values per second for individual hosts. Specifically this would be VERY helpful for seeing when trapper items are receiving far more data than expected.
For instance, lets say that my global NVPS suddenly jumps by 200, but I didn't add any new hosts and the global number of items has stayed relatively flat. How do I identify which host is causing the sudden jump? With small numbers of hosts/items it may only be somewhat easier to find the issue. I have 12k+ hosts and 1 million items. This is almost an impossibility to find.
Debug logs don't help here because with so many hosts/items it is incredibly difficult to spot an increase in data from a single host. With an NVPS counter per host it would be easy to create trigger that will alert me when hosts suddenly begin sending in more data than expected.