jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Deepak Shetty <shet...@gmail.com>
Subject Re: If i do a manual execution by adding extended status bar to firefox
Date Fri, 30 Sep 2011 16:55:00 GMT
how have you calculated it manually (the last row value is not much useful
anyway)?
If Request1  = 10 samples each 10 secs therefore avg =10 seconds
and request2 = 1 sample , 1 second therefore avg = 1 second
then overall average = (10*10 + 1*1)seconds/(10+1) samples.



regards
deepak


On Thu, Sep 29, 2011 at 11:50 PM, sprasad <sprasad.sadu@gmail.com> wrote:

> Hi Oliver,
>
> I have a small () question for you -
>
> I am running some tests and presenting the aggregate report - in this,
> Average Response Time, which is the second column in agg. report table is
> the "average of set of response times", but the the value displayed in the
> last row - Total - and under Average column is not matching up when i do
> the
> average of response times manually.
>
> For ex, have a look at the below table: I have taken Average in seconds.
> The
> Average *0.66* in the last row is not matching up when i do the average of
> the response time manually. Also, Min and Max represent the min. response
> time and max. response time for that particular label (navigation) - right
> ?
>
>
> *Label*
>
> *Samples*
>
> *Average (s)*
>
> *Median*
>
> *90%Line*
>
> *Min*
>
> *Max*
>
> *Error%*
>
> *Throughput*
>
> *KB/Sec*
>
> URL
>
> 75
>
> 1.00
>
> 749
>
> 3019
>
> 27
>
> 6035
>
> 0
>
> 8.39
>
> 164.26
>
> 1
>
> 50
>
> 0.63
>
> 34
>
> 2978
>
> 20
>
> 6031
>
> 0
>
> 7.38
>
> 53.07
>
> Login Page
>
> 25
>
> 0.14
>
> 17
>
> 33
>
> 12
>
> 3009
>
> 0
>
> 3.74
>
> 25.02
>
> 2
>
> 25
>
> 0.13
>
> 9
>
> 16
>
> 7
>
> 2939
>
> 0
>
> 3.74
>
> 5.69
>
> 3
>
> 25
>
> 1.10
>
> 864
>
> 1120
>
> 306
>
> 4106
>
> 0
>
> 3.36
>
> 146.53
>
> 4 Home Page
>
> 25
>
> 0.87
>
> 957
>
> 1028
>
> 234
>
> 1038
>
> 0
>
> 2.78
>
> 121.36
>
> 5
>
> 25
>
> 0.05
>
> 52
>
> 59
>
> 30
>
> 77
>
> 0
>
> 2.86
>
> 7.46
>
> *TOTAL*
>
> *250*
>
> *0.66*
>
> *58*
>
> *2961*
>
> *7*
>
> *6035*
>
> *0*
>
> *24.94*
>
> *427.27*
>
>
> Please clear my doubts.
>
> Thanks,
> Sd
>
>
> On Sat, Aug 20, 2011 at 5:11 PM, Oliver Lloyd [via JMeter] <
> ml-node+4718262-1402807349-230167@n5.nabble.com> wrote:
>
> > The problem here is you are trying to do something that is essentially
> not
> > possible and, crucially, not useful.
> >
> > JMeter is not a browser so it will never give you the exact same response
> > times for rendering a page as you see in your browser. Indeed, different
> > browsers themselves will give you different page load times so there's
> > really no point worrying about the difference.
> >
> > What you really need to do is work out your objectives - what are you
> > trying to do? What are your requirements for testing? You should qualify
> in
> > verbal terms what you what to achieve and then prove the same using your
> > tests. Then you should quantify these statements to create actual targets
> to
> > aim for, giving you pass / fail criteria. Having objectives is especially
> > important when thinking about 'rendering' the full page because this
> > experience is effected by multiple things and needs to be understood and
> > planned for. For example, do you even want to simulate page resources?
> Are
> > you using a CDN? Are there any other types of caches that you need to
> take
> > account of when designing your load?
> >
> > Based on the requirements you then go away and design a test that meets
> > them. If page rendering is in scope, if you want to validate the client
> side
> > performance, then this is more of a functional activity - you cannot use
> a
> > tool like JMeter to test this. In fact you don't need a 'tool' to do this
> at
> > all, just a mouse and keyboard will suffice. Client side performance is
> > crucial - there are lots of things to be considered and tuned for - but
> it
> > is not really relevant in this forum.
> >
> > So, in short, page load times and server response are part of the same
> user
> > experience but need to be approached in different manners - this is an
> age
> > old principle of performance testing which is routinely misunderstood.
> >
> > If you can't get your head around this then these days you can solve the
> > problem by paying an external company to run tests where each script runs
> in
> > it's own browser (the Cloud makes this possible) or you can go down the
> > hybrid path and use another (rather well known) load test tool that gives
> > you GUI like scripts that almost run full browsers. If you don't have the
> > budget for this then JMeter offers an excellent solution, however without
> a
> > basic foundation knowledge of computing you're very unlikely to do a good
> > job.
> >
> > ------------------------------
> >  If you reply to this email, your message will be added to the discussion
> > below:
> >
> >
> http://jmeter.512774.n5.nabble.com/If-i-do-a-manual-execution-by-adding-extended-status-bar-to-firefox-tp4717698p4718262.html
> >  To unsubscribe from If i do a manual execution by adding extended status
> > bar to firefox, click here<
> http://jmeter.512774.n5.nabble.com/template/NamlServlet.jtp?macro=unsubscribe_by_code&node=4717698&code=c3ByYXNhZC5zYWR1QGdtYWlsLmNvbXw0NzE3Njk4fC0xNDM4OTcyNzI4
> >.
> >
> >
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/If-i-do-a-manual-execution-by-adding-extended-status-bar-to-firefox-tp4717698p4856044.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>

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