Just out of curiousity, what particular performance data are you looking for? We've gone down the perfstat gather path many times and honestly we've only found it useful when we're doing a correlation analysis of something like an Oracle AWR output against the NetApp perfstat. If we're just digging into how things are performing, we've actually had more luck using OCPM to tell us where our high water marks are (in terms of IOPs/latency/etc) and then just do a combination of sysstat and statit commands during the problem periods. Our most recent case involved gathering lots and lots of perfstats (along with various statit gathers) but the TSE came back and used the statit data to sleuth out what the issue was (long running raid-scrubs in that case).
Hope that helps - if you've got some specific performance issue examples you can provide maybe that would help folks give feedback on where to look (based on similar experiences).