forum.e-valid.com

In your server loading runs how do I get more detailed timing out of a test scenario than just the total time of playback? Basically, I need to know how long a section in the middle of my script takes -- that is the one where the backend SQL server performance causes the most problems.

engpro wrote:In your server loading runs how do I get more detailed timing out of a test scenario than just the total time of playback? Basically, I need to know how long a section in the middle of my script takes -- that is the one where the backend SQL server performance causes the most problems.

You're right: the standard LoadTest reporting is for the time the whole script takes to execute. As server load increases you would expect this total time to increase and the LoadTest Monitor reports min/average/maximum times for each "track" (i.e. defined user type) that you are running.

To extract more detail you probably want to modify your script to include some SaveRecord commands that write out intermediate times into a local file. This command includes a way to timestamp the time the record is written, so you can correlate imposed server load (as a function of time) with the corresponding increase (or decrease) in the intermediate timings.

It's worth noting that multiple Browser Users (eValid instances) can all write to the same file; the Windows file system keeps the records from getting confused or interposed.

We've used this technique many times before and it works fine.

The eValid Tesm

eValid wrote:

engpro wrote:In your server loading runs how do I get more detailed timing out of a test scenario than just the total time of playback? Basically, I need to know how long a section in the middle of my script takes -- that is the one where the backend SQL server performance causes the most problems.

You're right: the standard LoadTest reporting is for the time the whole script takes to execute. As server load increases you would expect this total time to increase and the LoadTest Monitor reports min/average/maximum times for each "track" (i.e. defined user type) that you are running.

To extract more detail you probably want to modify your script to include some SaveRecord commands that write out intermediate times into a local file. This command includes a way to timestamp the time the record is written, so you can correlate imposed server load (as a function of time) with the corresponding increase (or decrease) in the intermediate timings.

It's worth noting that multiple Browser Users (eValid instances) can all write to the same file; the Windows file system keeps the records from getting confused or interposed.

We've used this technique many times before and it works fine.

The eValid Tesm

Can you provide some sample code and more detailed steps how this is done?

Thanks for requesting a sample code and detailed steps.

In situations like this we request that you contact our technical support group directly for the details you need.

When you contact them please provide (1) a copy of the script that you are trying to modify, (2) a description of what modifications you want to make, and (3) any other operational details that you think will be of assistance.

_______________
The eValid Team