Since response time variance is not always in a linear relation with
concurrency, I can't say that the response time with 1 request will be the
same as 18000 requests in 1 hour or 5 requests/second. Another way to do
this is using a Runtime Controller, with 1 hour run time, a global
counter expression, for example ${__counter(globalCounter,FALSE)}, and an
Action Test sampler inside an If Controller, to stop the test when it sums
18000 requests.
Thread Group ( 5 threads, 0 ramp-up, 1 loop, delayed threads checked)
- Runtime Controller ( 3600 seconds )
- WebService sampler ( include the __counter function here or add a
BeanShell sampler to execute the __counter function )
- If Controller ( ${globalCounter} > 18000 )
- Test Action ( All Threads, Stop Now )
Hope it helps.
Flávio Cysne
2012/10/23 Deepak Shetty <[email protected]>
> the kirk pepperdine technique :) if you need a constant input rate
> irrespective of the response time.
> I guess it depends on your response time - suppose the response time of the
> webservice is under (3600/18000) 0.2 seconds you haven't then introduced
> any degree of concurrency. I usually think of these in terms of concurrency
> rather than the number of requests at the server.
>
> regards
> deepak
>
> On Tue, Oct 23, 2012 at 5:03 AM, Flavio Cysne <[email protected]>
> wrote:
>
> > What do you say about 18000 threads, 3600 seconds ramp-up, 1 loop and
> > delayed threads (must use JMeter 2.8, awesome) ?
> >
> > 2012/10/22 Deepak Shetty <[email protected]>
> >
> > > Hi
> > >
> > > a. If you choose duration use a constant throughput timer to specify
> the
> > > rate of requests.
> > > b) If you want the test to finish as quickly as it can then just
> specify
> > > the number of loops and remove the duration
> > >
> > > dont specify both
> > > regards
> > > deepak
> > >
> > > On Mon, Oct 22, 2012 at 4:55 PM, Kouta Osabe <[email protected]
> > > >wrote:
> > >
> > > > hello,deepak
> > > >
> > > > I ment specfing loop count on the thread group and also gave a
> > scheduler
> > > > setting - one hour.
> > > >
> > > > regards,
> > > > kota
> > > > 2012/10/23 6:55 "Deepak Shetty" <[email protected]>:
> > > >
> > > > > when you say 60 loops did you add a controller or did you specify
> > loop
> > > > > count on the thread group.
> > > > > Did you also give a scheduler setting?
> > > > >
> > > > > regards
> > > > > deepak
> > > > >
> > > > > On Mon, Oct 22, 2012 at 12:26 PM, Kouta Osabe <
> > > [email protected]
> > > > > >wrote:
> > > > >
> > > > > > Hello All,
> > > > > >
> > > > > > I want to just 18,000 requests to api server for an hour.
> > > > > >
> > > > > > so i configured 300 threads, 60 seconds ramp-up, 60 loops.
> > > > > >
> > > > > > i expected just 18,000 requests sended to api server, but over
> > > 18,000
> > > > > > requests sended to api.
> > > > > >
> > > > > > Can someone please help me how i send just request to api server?
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > Kota.
> > > > > >
> > > > > >
> > ---------------------------------------------------------------------
> > > > > > To unsubscribe, e-mail: [email protected]
> > > > > > For additional commands, e-mail: [email protected]
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
