Thank you all for the responses. The simple data writer and transaction controller gave me what I needed.
After analyzing a few tests, I notice that jmeter times are slower than grinder times (we are leaving grinder). EG: a logon logoff took 1 second in jmeter, but 1/2 second in grinder. Will actions not relevant to the test increase Jmeter reported times? Actions such as post processors, beanshell, config elements and controllers, user defined vars and parameters : do they add to the test time? Also is the test time reported the time it took to send and receive the full response from the host? OK. thank you again. -------------------------------------------- On Wed, 8/27/14, Jeff Ohrstrom <[email protected]> wrote: Subject: Re: Jmeter reporting To: "JMeter Users List" <[email protected]> Date: Wednesday, August 27, 2014, 6:49 PM Also you could use a transaction controller as a parent element and group all 3 transactions. That way, in the csv file you have all 3 times grouped as 1 transaction. Then you can compare that transaction (the combination of all 3 subtransactions) across tests. It would look something like this: +--EntireUserTransaction --Login --Transaction --Logoff That way every 'EntireUserTransaction' you see in the result file is the user's experience. Then you can do analysis on that transaction and not have to worry about which thread made which transaction. On Mon, 2014-08-25 at 21:00 +0000, Jim Brannan wrote: > If you use the simple data writer and save the data to a csv file, you will see the data can include the thread name, which will end in like 1-1, 1-2, 1-3... The -1, -2, -3, etc are the individual users. So you can easily do a pivot summary with this. I think someone needs to write a save-to-sql listener!!! For now I have a tear-down thread calling an EXE that reads the saved data file and inserts the data (improved for reporting) into an SQL table. 'Opportunity' with performance testing is that is makes a LOT of data. > > Jim > -----Original Message----- > From: rxfillpharm [mailto:[email protected]] > Sent: Monday, August 25, 2014 12:50 PM > To: [email protected] > Subject: Jmeter reporting > > Hi Group, > I posted a question some days ago and guess it was not clear what I need. > New to performance testing and jmeter. I have a bunch of test plans that work fine. > > > My test logs on, runs a small query, then logs off. > That takes @ 5 seconds as per my manual test and the jmeter test. > > When I increase threads (users) to 10, I want to know with all users running concurrently, > how long each individual user takes to complete the test (log on, run query, log off). > There should not be that much of a different. > > When I increase threads to 100, I should see a difference in the total time per users. > Increase to 500 users, I should see the timer for each user take longer still. > > Is there a listener in jmeter that can tell me that, ie time per user. I'm thinking one > row of raw data for each thread. I am not interested in the per http call speed. > > Thank you and I hope that my question is clear(er). > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
