Unanswered question

How does Neoload measure request response time?

I would like to ask sincerely how Neoload measure request response time or how request response time worked. As shown in the attachment picture I have created test for several transactions like "login in", "project loading"(loading into a list of project), "project inside"(loading into the project details). However I found it super weird that the time I recorded does not reflect the "real situation" when I try the user path I recorded manually(the recorded time and real time is different). For example for "project inside", I don't feel like I need 30 sec to load the whole page. This is not a special case, I have tried to perform the load test with the same user path(recording) few times and I get 3x seconds for this action every time.

I would mention what I understand first to let you know what I really don't understand.

1. From my understanding, web performance time( or I would say it as the real time" is equal to page rendering time + request response time and Neoload is only responsible for measuring the latter part. That means the request response time Neoload recorded should be less than the real time.

2. I know the definition of request response time is something like the time when the first Request is sent and the time of the last byte of the response of the last Request. However, the problem now is every request response time recorded seemed larger than what the actual time are, that makes no sense. So, I would like to know the reason behind for this record time > actual time problem (Perhaps although we(users) can see the page completed loading, but there are still responses not yet to come?)

Thanks for your time. Any comments are appreciated.

chun tat K.
chun tat K.

chun tat K.

7 / 100


From your screen shot we can see that the request responsible of the high response time is the /Clarizen/push request.

According to the name of the request it might be a PUSH mechanism that is used by your application and not recognized by NeoLoad as PUSH. In other words maybe that request should not be replayed in the main thread of your script but in parallel of your script.

If you use the developer tools of your browser while browsing your application do you see that request? If yes do you get a response right away from the server or is it also a high response time like in NeoLoad?

Does this request appears once or multiple times in your script/browser?

Also have you checked in the check virtual user that you get the same response from the server during recording to make sure that there's no error?