> Things pertaining to the performance of particular webpages and such are extremely difficult to tease out of the data in a meaningful way.

It seems to me that you could do this "well enough" if you put each page in its own process and let the OS tell you which are burning the CPU. There could/would be edge-cases where a page causing issues would not appear like it was causing issues, looking at it from that level, but it would sure as hell be a good start.