How does this work when using transaction controllers?
For example the following testplan: +Thread group: 1 user every 5 seconds +++Transaction Controller ++++++HTTP Request1 --> Return code 302 redirect ++++++HTTP Request2 --> Return code 200 with 18 embedded resources
Will this generate 0,2 hits per second or 4 hits per second?
To continue the previous question, the "problem" is not related to the transaction controller, since the same behaviour can be reproduced with a single HTTP request sampler that has the "retrieve all embedded resources" flag checked (even without a 30x redirection).
Since the "view test result" listener can understand that the main request (the url provided in the sampler) has generated some other child request, can't the Hits/sec listener do this? I think that this could be a more precise count on generated hits.
I have 2000 users with a thread delay of 1 sec (set in Constant Timer), i expect my Hits/sec to be around 2000, but i see the total number of Hits/sec (as per your graph), as 56 hits/sec only...and average response time is : 0.01312 sec. Not sure what is going on here. Could you someone please explain?
Hi, it shows your test produces 56 hits per seconds. If it is not what you expected, it is maybe because you are not interpreting correctly how JMeter thread work. Are you sure you have 2000 running at the same time? JMeter cannot handle that many concurrent threads...
I currently use Thread Group, to assign my user count and ramp up. If i Number of users as: 1000, and ramp up time of 500 secs, isn't it that after 500 secs, i have 1000 active threads shooting requests? I have a constant timer (thread delay) set for 2 ms, then shouldn't i have at least 500 hits/sec...
Comment
by project member
a...@apc.kg,
Apr 14, 2011
Remember that each thread won't send next request until previous response received. So if you have constant delay of 5ms and server response time of 150ms then you'll have maximum of 1000/155 = 6.45 requests per second for each thread.
I desperately need this one. Is it already implemented (cannot find it in the Listeners list)?
Hi, Alexander, It is not released yet, but you can have it in beta. Ask if I should upload a beta snapshot for you.
That would be great! Thank you in advance.
Ok, here is it: http://code.google.com/p/jmeter-plugins/downloads/detail?name=JMeterPlugins-0.3.1-beta1.zip
Thanks for the link. Unfortunately I cannot not try it out, because the JAR file is not included.
There's really was an issue with JAR in archive. Try this link: http://code.google.com/p/jmeter-plugins/downloads/detail?name=JMeterPlugins-0.3.1-beta2.zip
премного благодарен ;)
What's the difference between this graph and Transaction per second?
Let's take an example: 3 server request at time t0, 1 response t0+1, 2n response t0+2, 3rd response t0+3 Hits/sec = 3 Transactions/sec = 1 over 3 sec
How does this work when using transaction controllers?
For example the following testplan: +Thread group: 1 user every 5 seconds +++Transaction Controller ++++++HTTP Request1 --> Return code 302 redirect ++++++HTTP Request2 --> Return code 200 with 18 embedded resources
Will this generate 0,2 hits per second or 4 hits per second?
Hi Sander,
The transaction controller hides what is inside, it counts as 1 hit.
Hi all and compliments for this plugin! :)
To continue the previous question, the "problem" is not related to the transaction controller, since the same behaviour can be reproduced with a single HTTP request sampler that has the "retrieve all embedded resources" flag checked (even without a 30x redirection).
Since the "view test result" listener can understand that the main request (the url provided in the sampler) has generated some other child request, can't the Hits/sec listener do this? I think that this could be a more precise count on generated hits.
Kind regards ;)
Hi,
I was wrong, the transaction controler do not hide what is inside. It is now fixed for both cases on hit/sec, latency and resp code charts.
Hi Stephane,
great work! When this fix will be available for download?
Ps: sorry for posting here, but I think the question can be interesting for some other people than me... ;)
Hi, Current snapshot available here: http://code.google.com/p/jmeter-plugins/downloads/list?can=1
Hi,
Just want to know whether this is same as number of requests/sec?
Hi, yes a hit is a request to the server.
I have 2000 users with a thread delay of 1 sec (set in Constant Timer), i expect my Hits/sec to be around 2000, but i see the total number of Hits/sec (as per your graph), as 56 hits/sec only...and average response time is : 0.01312 sec. Not sure what is going on here. Could you someone please explain?
Hi, it shows your test produces 56 hits per seconds. If it is not what you expected, it is maybe because you are not interpreting correctly how JMeter thread work. Are you sure you have 2000 running at the same time? JMeter cannot handle that many concurrent threads...
I want generate around at least 600 hits/sec on server. Do you think i can use Ultimate Thread Group for this?
I currently use Thread Group, to assign my user count and ramp up. If i Number of users as: 1000, and ramp up time of 500 secs, isn't it that after 500 secs, i have 1000 active threads shooting requests? I have a constant timer (thread delay) set for 2 ms, then shouldn't i have at least 500 hits/sec...
Remember that each thread won't send next request until previous response received. So if you have constant delay of 5ms and server response time of 150ms then you'll have maximum of 1000/155 = 6.45 requests per second for each thread.
Maybe you'd find useful the Shaper.
Thank you so much for your help..!! :)
How can I get the average Hits/Sec ??
The graph displays after run the scenario or read the JTL file is quite different, why it happens?