jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sebb <seb...@gmail.com>
Subject Re: JMeter Latency Showing incorrect ?
Date Mon, 11 Jan 2010 00:23:07 GMT
On 10/01/2010, Deepak Shetty <shettyd@gmail.com> wrote:
> hi
>  so what do you think is the cause behind latency being reported as 0(is it
>  the granularity)?

Yes

> a quick glance through the code showed that latencyEnd is
>  not called if the content length isn't 0 and some property
>  httpsampler.obey_contentlength is set to true  , but i dont think this
>  property is set by default

Oops!

No, that is not the default - I've run a test that shows non-zero
latency for 200 and 304 status returns.

>
>  regards
>
> deepak
>
>
>  On Sun, Jan 10, 2010 at 3:51 AM, sebb <sebbaz@gmail.com> wrote:
>
>  > On 10/01/2010, Deepak Shetty <shettyd@gmail.com> wrote:
>  > > I guess someone familiar with the code base would have to reply , I guess
>  > >  latency may not be calculated if the status code is not in 2xx.
>  >
>  > Yes, it is calculated for both HTTP Samplers.
>  >
>  > >  Alternately its possible that the latency is an extremely small value (<
>  > 1ms
>  > >  which is the granularity for older JVMs) since the server is simply
>  > >  returning you a status which is why you see zero(means less than 1 ms).
>  >
>  > Timer granularity may be a lot higher than 1ms.
>  >
>  > However, if the test plan is being run using Java 1.5 or later, the
>  > code will automatically use System.nanoTime() for elapsed time
>  > measurement, which should have much lower granularity.
>  >
>  > >  regards
>  > >
>  > > deepak
>  > >
>  > >
>  > >  On Sat, Jan 9, 2010 at 10:37 PM, Jatin <daveyjatin@gmail.com> wrote:
>  > >
>  > >  > You are correct deepak , I had redirects in the first sampler.
>  > Observerd
>  > >  > the HTTP redirection status code in the view results listener. After
>  > >  > removing that sampler i had latency being shown correctly.
>  > >  >
>  > >  > Again , why is the latency shown 0 for redirects as even the redirects
>  > >  > traverse across the internet and it takes time for Jmeter to get them
>  > >  > process them.
>  > >  >
>  > >  >
>  > >  > Thanks
>  > >  > Jatin
>  > >  >
>  > >  > Deepak Shetty wrote:
>  > >  >
>  > >  >> Hi
>  > >  >> I tested against www.google.com and I dont see a zero latency. did
>  > you
>  > >  >> verify with View Results Tree that you are indeed getting Google's
>  > home
>  > >  >> page
>  > >  >> (no redirects/errors or anything)? are you seeing this behavior
for
>  > all
>  > >  >> external sites?
>  > >  >>
>  > >  >> The elapsed time = latency +( time from 1st byte to last byte
>  > received).
>  > >  >> Any combination of value is possible , with elapsed time always
>  > greater
>  > >  >> than
>  > >  >> latency.
>  > >  >>
>  > >  >> regards
>  > >  >> deepak
>  > >  >>
>  > >  >>
>  > >  >> On Sat, Jan 9, 2010 at 10:11 PM, Jatin <daveyjatin@gmail.com>
wrote:
>  > >  >>
>  > >  >>
>  > >  >>
>  > >  >>> I was testing from my home laptop. My internet speed is 256
kbps. I
>  > >  >>> recorded the home page of google.com and was replaying it which
had
>  > 9
>  > >  >>> samplers to access all the contents of the home page. I am confused
>  > on
>  > >  >>> one
>  > >  >>> point , the Elapsed time is mostly above 400ms for the first
sampler
>  > >  >>> where
>  > >  >>> as the latency for this sampler is always 0 no matter how many
times
>  > i
>  > >  >>> loop.
>  > >  >>> Both Elapsed time and latency have one thing in common which
is that
>  > both
>  > >  >>> take into account the delay happening because of the response
>  > getting
>  > >  >>> from
>  > >  >>> across the internet, then how come latency is 0 all the time
while
>  > the
>  > >  >>> elapsed time is more than 400ms for most of the time.
>  > >  >>>
>  > >  >>> Thanks
>  > >  >>> Jatin
>  > >  >>>
>  > >  >>>
>  > >  >>> Deepak Shetty wrote:
>  > >  >>>
>  > >  >>>
>  > >  >>>
>  > >  >>>> if you are testing this out within your intranet , its possible
you
>  > >  >>>> could
>  > >  >>>> see extremely low or zero values.
>  > >  >>>> regards
>  > >  >>>> deepak
>  > >  >>>>
>  > >  >>>> On Sat, Jan 9, 2010 at 2:59 AM, Jatin <daveyjatin@gmail.com>
>  > wrote:
>  > >  >>>>
>  > >  >>>>
>  > >  >>>>
>  > >  >>>>
>  > >  >>>>
>  > >  >>>>> Hi all
>  > >  >>>>>
>  > >  >>>>> I had a recorded script that will access the home page
of a site.
>  > The
>  > >  >>>>> listener that i was using was a graph results listener.
I wanted
>  > to
>  > >  >>>>> save
>  > >  >>>>> the
>  > >  >>>>> results in a csv file for only two variables namely:
latency and
>  > >  >>>>> Elapsed
>  > >  >>>>> time. So in the configure dialog only selected these
two
>  > parameters and
>  > >  >>>>> obtained a csv file.
>  > >  >>>>>
>  > >  >>>>> My test plan organization was comprised of 9 HTTP samplers
in one
>  > >  >>>>> thread
>  > >  >>>>> group with all the samplers put into a loop controller
looping
>  > forever.
>  > >  >>>>> For
>  > >  >>>>> all the tests that were looping in the controller i
noticed that
>  > for
>  > >  >>>>> the
>  > >  >>>>> first sampler the latency was always 0 even though the
elapsed
>  > time was
>  > >  >>>>> around 400ms but not less than that. Is the latency
value correct
>  > since
>  > >  >>>>> for
>  > >  >>>>> both these parameters the response has to be fetched
from the
>  > server
>  > >  >>>>> and
>  > >  >>>>> there could be delay in receiving the response. Please
let me know
>  > if i
>  > >  >>>>> am
>  > >  >>>>> missing something here.
>  > >  >>>>>
>  > >  >>>>> Some test run samples are as below :
>  > >  >>>>>
>  > >  >>>>> 459,0
>  > >  >>>>> 140,137
>  > >  >>>>> 63,40
>  > >  >>>>> 179,134
>  > >  >>>>> 137,137
>  > >  >>>>> 120,120
>  > >  >>>>> 53,39
>  > >  >>>>> 40,40
>  > >  >>>>> 113,113
>  > >  >>>>> 275,0
>  > >  >>>>> 157,154
>  > >  >>>>> 62,40
>  > >  >>>>> 540,144
>  > >  >>>>> 149,148
>  > >  >>>>> 33,33
>  > >  >>>>> 53,38
>  > >  >>>>> 37,37
>  > >  >>>>> 113,113
>  > >  >>>>> 470,0
>  > >  >>>>> 159,155
>  > >  >>>>> 62,39
>  > >  >>>>> 178,133
>  > >  >>>>> 176,176
>  > >  >>>>> 120,120
>  > >  >>>>> 54,39
>  > >  >>>>> 38,37
>  > >  >>>>> 113,113
>  > >  >>>>>
>  > >  >>>>>
>  > >  >>>>> Thanks
>  > >  >>>>> Jatin
>  > >  >>>>>
>  > >  >>>>>
>  > ---------------------------------------------------------------------
>  > >  >>>>> To unsubscribe, e-mail:
>  > jmeter-user-unsubscribe@jakarta.apache.org
>  > >  >>>>> For additional commands, e-mail:
>  > jmeter-user-help@jakarta.apache.org
>  > >  >>>>>
>  > >  >>>>>
>  > >  >>>>>
>  > >  >>>>>
>  > >  >>>>>
>  > >  >>>>>
>  > >  >>>>
>  > >  >>>>
>  > >  >>>>
>  > >  >>>
>  > ---------------------------------------------------------------------
>  > >  >>> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
>  > >  >>> For additional commands, e-mail:
>  > jmeter-user-help@jakarta.apache.org
>  > >  >>>
>  > >  >>>
>  > >  >>>
>  > >  >>>
>  > >  >>
>  > >  >>
>  > >  >>
>  > >  >
>  > >  >
>  > >  > ---------------------------------------------------------------------
>  > >  > To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
>  > >  > For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>  > >  >
>  > >  >
>  > >
>  >
>  > ---------------------------------------------------------------------
>  > To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
>  > For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>  >
>  >
>

---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org


Mime
View raw message