Improve p2 download stat collection

manderse@redhat.com said:

We shouldn't track every feature - should review the list and also try add some more context than just the feature name if possible.

Questions to discuss in order to flesh out this request:

a) if not all, then which features? Where do we get the list? Do we need to parse the JBT aggregate site's category.xml and pull out all features from a particular category, such as the Core (Abridged) category?

b) what context do you want instead?