hi im just trying to figure out the results from the load test that i have conducted.
can anyone please help me understand the rationale behind this:
one of the goals of the load test i have made is to get the average response time for the overall test. i have set the parameter to fail this goal if the value is greater than 2000ms. to which the goal failed because the max value reached is 11249ms
but on the table produced after the test, the average time is at 222ms only. the max is at 11249. so if the average response time is only at 222ms, and at one point there was a page that took 11249ms to load how does the application fair overall during the load testing. its quite confusing since the term "average response time (ms) for overall test" is used.
can anyone please help me understand the rationale behind this:
one of the goals of the load test i have made is to get the average response time for the overall test. i have set the parameter to fail this goal if the value is greater than 2000ms. to which the goal failed because the max value reached is 11249ms
but on the table produced after the test, the average time is at 222ms only. the max is at 11249. so if the average response time is only at 222ms, and at one point there was a page that took 11249ms to load how does the application fair overall during the load testing. its quite confusing since the term "average response time (ms) for overall test" is used.
5 Answers, 1 is accepted
0
Hello Niccolo,
Here is how this is working. When you setting up a load test you select a sampling rate (refer to the screenshot attached). So like shown in the screenshot it will take sample on each 30 seconds. So for 10 minutes test it will take 20 samples. So the 1st sample taken may be 100ms, the second sample to be 105 ms, the third 102ms etc. but at some point one of the samples can be 10 000ms for some reason, and this will be in the middle of the test. In this case the goal set will fail.
The average of all samples can be 222 but that will be the average of all 20 samples, taken during the entire test. This value is formed at the end of the entire test, so if this was the value that is supposed to fail the test it does not make sense since the test is already finished.
So the goal is set for those samples rates taken during the test.
Hope this is now a bit more clear for you.
Regards,
Ivaylo
Telerik
Here is how this is working. When you setting up a load test you select a sampling rate (refer to the screenshot attached). So like shown in the screenshot it will take sample on each 30 seconds. So for 10 minutes test it will take 20 samples. So the 1st sample taken may be 100ms, the second sample to be 105 ms, the third 102ms etc. but at some point one of the samples can be 10 000ms for some reason, and this will be in the middle of the test. In this case the goal set will fail.
The average of all samples can be 222 but that will be the average of all 20 samples, taken during the entire test. This value is formed at the end of the entire test, so if this was the value that is supposed to fail the test it does not make sense since the test is already finished.
So the goal is set for those samples rates taken during the test.
Hope this is now a bit more clear for you.
Regards,
Ivaylo
Telerik
0
Nick
Top achievements
Rank 1
Rank 1
answered on 08 Oct 2014, 10:31 AM
Hi Ivaylo,
Kindly clear this one out for me, if the goal is set for the sample rates and one sample takes more time than the others then only one part of the overall samples failed?
sample rates aside, and considering the overall performance of our application based on the the average response time during the load test is it safe to assume that it actually passed?
thank you.
Kindly clear this one out for me, if the goal is set for the sample rates and one sample takes more time than the others then only one part of the overall samples failed?
sample rates aside, and considering the overall performance of our application based on the the average response time during the load test is it safe to assume that it actually passed?
thank you.
0
Hello Niccolo,
Yes you are nearly right. The samples are on equal intervals, and in one of the intervals (one of the samples) the average response time for particular reason has been 11249 for some reason. But yes most of the time the average response time was pretty good as you can see its value of 222.
Regards,
Ivaylo
Telerik
Yes you are nearly right. The samples are on equal intervals, and in one of the intervals (one of the samples) the average response time for particular reason has been 11249 for some reason. But yes most of the time the average response time was pretty good as you can see its value of 222.
Regards,
Ivaylo
Telerik
0
Nick
Top achievements
Rank 1
Rank 1
answered on 09 Oct 2014, 03:15 PM
Thank you Ivaylo! Response appreciated.
0
Hello,
You are most welcome.
I am closing this ticket. In case you need further assistance please do not hesitate to contact us back.
Regards,
Ivaylo
Telerik
You are most welcome.
I am closing this ticket. In case you need further assistance please do not hesitate to contact us back.
Regards,
Ivaylo
Telerik