jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Deepak Shetty <shet...@gmail.com>
Subject Re: Odd problem with performance of JMeter
Date Wed, 22 Jun 2011 21:40:09 GMT
>Try putting
>it as a child of the main request in your journey - sounds like you only
>have one so that's easy. Then, like that, each thread will only run X times
>per minute where X is the number you give it.

The non intutive part refers to the way it calculates when it needs to run.
if you ask for 10 requests per minute and it takes 1 second for each
request, the sample runs for 10 times for 10 seconds and then waits 50
seconds (if I remember correctly) whereas most would expect it to run once
per six seconds on average.

regards
deepak

On Wed, Jun 22, 2011 at 2:18 PM, Oliver Lloyd <oliver_lloyd@hotmail.com>wrote:

> Just a note on the Constant Throughput business. It's not intuitive, it's
> true. But if you select - 'this active thread only' and also remember that
> it will apply to all threads in scope then it's actually a really solid
> little workhorse - it's one of the things I like best about JM. Try putting
> it as a child of the main request in your journey - sounds like you only
> have one so that's easy. Then, like that, each thread will only run X times
> per minute where X is the number you give it.
>
> Say you have 100 threads and a value of 30 for the CTT, that will generate
> 50 requests a second. Then, so long as your response times remain below 1
> second your test will hold this rate of requests and you can ascertain how
> your system performs and if it will maintain this load without degrading
> (memory leaks etc). If response time is or rises above 1 second then the
> load will not be achieved.
>
> You could equally have 200 threads and X as 15 in the CTT - you'd get the
> same throughput but now your requests would have a 2 second window to
> complete in.
>
> Note. Typically, this sort of load with the response body sizes you mention
> should be doable in the GUI, you could probably also use a few listeners,
> give it a whirl. (Notice the heavy use of conditional tense there though!)
>
> The CMD line option is brilliant , another reason I like JM, but Sebb is
> right (he's always right) you must have the gen summary results listener -
> you're blind otherwise.
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Odd-problem-with-performance-of-JMeter-tp4514951p4515526.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message