0

Having recorded a script in URL mode, created a workloadmodel that calculates the # pages per second I am expecting, and now wanting to monitor this graph (Pages Downloaded per Second) during the scenario run, I was tiped to enable this manually in the run time settings. So I did and the graph is displayed, but I do only get the AVG (Pages Downloaded per Second), not max/min ++.

Is this expected behavior of this graph?

Magnus Jensen
  • 905
  • 6
  • 36
  • 73

2 Answers2

1

As you are in URL mode each request is recorded, with multiple URL requests making up a page. Use transaction markers to mark the start and end of a page definition (without thinktime inside of the markers) and then watch the graph for transactions per second. As each transaction defines a page, then this should match to your need.

Note, transactions are logical, so you can have them nested, such as

start_transaction("Page");
start_transaction("Home_Page");
....
<logic to set status for pass|fail>
end_transaction("Home_Page",status_from_test);
end_transaction("Page",LR_PASS);

You could then just watch the Filtered graph for the "Page" transactions....

James Pulley
  • 5,606
  • 1
  • 14
  • 14
0

Another option to James suggestion is to not use URL mode, and actually load pages with resources on them. This will produce exact PAGES/SEC graphs.

This can be set in "recording options | recording | HTML-based script". Here you should also open the "HTML Advanced" dialog and select "A script containing explicit URL's only ...." - This will produce easier to read and handle scripts.

K.Sandell
  • 1,369
  • 10
  • 19
  • Ok, but I fond that recording the script in URL mode Solved a lot of issues with my RIA application. I assume recording the script in URL mode actually gives me something more than regular html mode, however I am not certain on what exactly causes this. I have red that URL mode is the preffered mode for RIA applications containing JS, Ajax etc being more of a thick client. What does URL mode provide over html mode given what I have just explained? – Magnus Jensen Mar 25 '13 at 20:13
  • You are right that in many cases the URL mode is better, but also much more demanding when it comes to correlation. Some JavaScript (Ajax mostly) does not get recorded with the HTML mode... It's always a trade off. I prefer the HTML mode if it works, and if only a few places need URL mode I re-record them in URL mode and just copy-paste the code.. – K.Sandell Mar 25 '13 at 20:22
  • 1
    Ok, Thank you. I found a good explaination online which states my current issue: recording in html mode Sometimes HTML based scripts cannot catch all the nuances neccessary to successfully simulate the end user. This usually is seen when application sensitive steps are erroneously found in EXTRARES sections or hidden in such a way that you need web_custom_request() or similar calls to actually simulate the real steps performed by the end user. This is my issue. Using html mode replay fails because of the things mentioned above, but with URL mode it works because the css, JS is recorded seperat – Magnus Jensen Mar 25 '13 at 20:26
  • I assume this is why URL mode suites me better in my case. However as you say, it might be a solution to record in html mode first, and then in URL mode, and finally paste in from the URL script to html script where it fails with the EXTRARES (within the html script) – Magnus Jensen Mar 25 '13 at 20:30