In a recent post I wrote that there are many ways to solve a problem.
In this post I will talk about the Perfmon and specifically about performance counters collection. Brent Ozar wrote a great article a while back on the Perfmon Best practices and in his article he suggests the use of perfmon logs and excel spreadsheets to collect and analyze SQL Server and system performance.
One way to solve a problem:
This method is great, however it has a few limitation and drawbacks:
- the user gets easily confused by the constantly changing UI of the perfmon with every OS release
- it is hard to align it all in excell – too many clicks
- it is not suitable for collection and analysis of multiple servers
- it is not suitable for analysis and comparison of multiple servers at different times
- and so on
There are a few benefits, though:
- no need for knowledge of command shell or SQL Server
- no need for knowledge of T-SQL
Another way to solve a problem:
Recently I blogged about another method for performance counters collection: Collecting Performance Data into a SQL Server Table.
Why is this method a bit better:
- after setting it up once and creating a template, the same solution can be used for analysis of multiple servers at different times
- the data import is create behind the scenes in a relational data store
- the solution is UI independent
- better performance and easy analysis
The drawbacks are:
- need to know command shell and SQL Server basics
- need to know T-SQL
- takes a bit more time to set up and develop, but the solution can be reused multiple times
It is up to you which method you will chose.
Feel free to contact me if you have questions.