When it comes to identifying and eliminating SQL Server performance issues, it is always better to be proactive than reactive. This article from Narayana Vyas Kondreddi shows you how to analyze the data gathered by Profiler to identify performance issues and also provides four generic stored procedures for analyzing Profiler output.
The idea is to run Profiler during a "load test" (or stress test) and analyze the Profiler output at the end of the load test to identify the longest running queries, the most CPU-intensive queries, stored procedure recompilations, errors like deadlocks, etc. For this purpose, Narayana uses a specific trace template that captures all the data necessary for the analysis.
The article continues at
Back to Database Journal Home