• raj more (4/16/2009)


    Great intro. I like it.

    I have seen tools like Intuit (now Quest S/w) Coefficient that do a phenomenal job of analyzing the output of a trace and posting the results in an extremely easy to use format.

    Can you post some useful trace templates that you use for Profiler 2008 and how to use them?

    Also, how do you baseline a server using Profiler and how can this info be best used when something starts acting up.

    Raj,

    Thanks for the kind words.

    I hadn't even considered posting templates mainly because templates I do create are usually just to add filters on database or application so I don't have to keep re-creating the filters, so they are specific to my environment. It's a great suggestion though.

    As far as baselining, that will be something I discuss in one of the later articles in the series. A quick idea is to to use the CPU, reads, and writes columns on the RPC:Completed, SP:Completed, and SQL:StmtCompleted events and see if there are changes to these as your data changes or as you make changes to queries. I've just started Grant Fritchey's book, SQL Server 2008 Query Performance Tuning Distilled, and I currently reading his chapter on baselining. He says, and I agree, you can't tell if what you are doing works if you don't have a baseline.