System Level Data

I was scanning Tom Kyte's blog and found a comment referring to Mogens Noorgard's blog We Do Not Use Blogs. I took a look and his latest discussions have concerned system wide performance and if Statspack is useful. There is a general consensus among the enlightened that Statspack, and system wide data in general, is not useful for performance tuning and actually hurts the effort to fix poorly performing jobs. I don't disagree, but I think we're missing the point. As I've pointed out for a long time now, Statspack or the data that it's based on, is not performance tuning data, it's capacity data. In other words, use system level data for capacity planning and workload measurement. It's not the case that it's useless. It is just misused! If you're really interested in this, see appsdba.com. By the way, I whole heartedly agree with using extended SQL trace data for performance tuning, or response time optimization.