Load Testing Scale

Hello Josh,

Are you referring to the Scale column as shown in the attached screen shot? These values are dynamically calculated by the UI as it's generating the graphs using the data pulled out of the database. It is auto-scaling the graphs of the data so that you get the best view of the data. Imagine a fixed scale where line 1 is .1-.5 and line 2 is 0-1000. If we used the same scale for all both lines then line 1 would be just a flat line. You would never see any changes over time of that data.

The data stored in the database always uses the same units. If you're pulling the data directly yourself then I don't really understand the problem. Can you clarify what challenge you are running into?

Greetings,
Cody
the Telerik team

Apologies, I think I misunderstood the metrics I'm talking about in particular - Time On Wire per Interval (sec) & Time To First Byte per Interval (sec). After a bit of looking, they don't appear in the Test Studio "Analyze" page but they do appear in the database.

So consider the following a separate issue.

I've attached an image displaying the problem I'm having. Is it just me or are these values too high? Considering the fact that I observed the test and there were no wait times this bad, is it possible this is just an anomaly? Or is it just a product of the way Test Studio gathers and averages the data?

Hi Josh,

Yes those values are much too high. There's a know bug where in certain circumstances the calculations were incorrect and the incorrect values got stored in the database. Our latest internal build, 2012.2.1527, contains a fix for these calculation errors. I recommend upgrading to this more recent version.

Greetings,
Cody
the Telerik team

If I upgrade to the latest internal build, will I also have to upgrade my Test Studio runtimes?

Hi Josh,

Yes you will. It's very important that all machines be at the same version of Test Studio that will be touching the same project. It would be OK to upgrade just one machine verify everything works just on it first before upgrading all the other machines. This is a very common practice.

Regards,
Cody
the Telerik team

Thanks for the assistance, after a universal upgrade the monitoring seems to be performing as intended.

Hello,

I am very happy to hear this is resolved! Thank you for the update.

Regards,
Cody
the Telerik team