Hi, I want to baseline a specific SQL statement to be able to tune performance and see impact of changes made to the SQL Statement. I have a Business Transaction filtered on the SQL Statement and in calculate results I have selected the response time of that SQL statement. I have added the Business Transaction to the application overview to begin baselining. But the thing is, if I understand this correctly, the baseline there is going to be for the response time of the entire PurePaths captured by the Business Transaction whereas I only want to baseline the response time of just the SQL statement within the PurePath. Is there a way to do that?
Assuming that baselining a specific SQL statement is not possible, what I'm planning alternatively is simply comparing the average over a time frame before a tweak to the SQL statement and right after to see if the response time has gone up, down, or remained unchanged.
Solved! Go to Solution.
Could you create a measure on the SQL statement then just graph it over time with the aggregation set to avg? I'm not sure the use case that would have a particular SQL statement changing all the time where baselining it would be a better idea then just making a measure on it. I realize this does not answer the question, just trying to see if there is an alterntative to accomplish what you need.
Shane and I discussed the use case a little further offline. But essentially the use case is that my customer has some specific SQL statements they want to fine tune, so they will make a tweak and I'll let them know how this impacted the response time of the statement.
There isn't a way to baseline a specific SQL query, so what I'm doing alternatively is simply comparing the average response time before and after a tweak to the SQL statement by using the response time measure of my Business Transaction. I'm using a meter chart just to make the comparison more straight forward.
My understanding is that the baselining is somewhat limited in that it only applies to the PurePath response time, failure rate, and throughput of business transactions (added to the applications overview) that are split by application and at most one other measure (with some exclusions such as agent). So while there may be some options such as what Shane mentions there won't be a way to baseline the SQL time (or any other result measure you may add) and that would be an enhancement request.