jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Maria Alejandra Trozzi <matro...@gmail.com>
Subject Re: Distribution Graph and Graph
Date Tue, 05 Jan 2010 12:26:16 GMT
You're right, I didn't realise of that... in the Aggregate Graph I see a bar
for each "Page" (request) and in the other graph I only see "times"....
alltogether
Thank you very much for the assistance!!
Ale.

2010/1/4 Deepak Shetty <shettyd@gmail.com>

> Ill have to verify but they would be the same only if you have exactly one
> sample(which may be requested N number of times).
>
> The Aggregate Graph plots the 90 % line per sample, so you would get a
> different 90% line for each sample . What I mean is say your tests make 3
> requests, request1, request2 and request3,(each requested say 100 times =
> 300 samples) the Aggregate graph would plot a 90% line considering each
> request at a time , i.e. 100 samples at a time i.e. you would have 3 bars.
> The Distribution graph on the other hand does not consider the sample and
> so
> would consider all 300 requests and calculate the 90th percentile which
> would only be the same in the special case that there is only one sampler.
> In all other cases they will probably be different.
>
> regards
> deepak
>
> On Mon, Jan 4, 2010 at 6:24 AM, Maria Alejandra Trozzi
>  <matrozzi@gmail.com>wrote:
>
> > I still have one question: the 90line in the graph doesn't have to be the
> > same value as in the other reports (Aggregate Report, or Aggregate Graph)
> > for 90line?
> > It hasn't the same value in my results... could you check in yours?? I
> > don't
> > have a place to put the files quickly....
> >
> > Thank you very much for your answer and time!!
> >
> >
> > 2009/12/29 Deepak Shetty <shettyd@gmail.com>
> >
> > > the mailing list blocks attachments, you must upload your files
> someplace
> > > and provide the links
> > >
> > > I havent used the distribution graph but from my reading of the manual
> > > Lets say you have a single request which has response times of
> > > 1,1,2,2,3,3,4,4,5,6 milliseconds - 10 samples
> > > then on the X axis you will plot a bar for each unique response time
> i.e.
> > 6
> > > bars (1,2,3,4,5,6 milliseconds)
> > > On the Y axis you will plot the percentage of times the response time
> > > occurs
> > > (1 = 20%,2=20%,3=20%,4=20%,5=10%, 6=10%)
> > > 50% threshold line will coincide with the bar graph for the 3 second
> > > response time
> > > 90% threshold line will coincide with the bar graph for the 5 second
> > > response time
> > >
> > >
> > >
> > > regards
> > > deepak
> > >
> > >
> > >
> > >
> > > On Tue, Dec 29, 2009 at 9:58 AM, Maria Alejandra Trozzi
> > > <matrozzi@gmail.com>wrote:
> > >
> > > > I'm getting some "strange" graphs in my tests....  Here I attach the
> > > > Distribution Graph for 20users and 50 users and the Graph of Response
> > > > Times... for 100 users.
> > > > I've read in the manual that a distribution graph like the one of 50
> > > users
> > > > with the 90line outside the graph may mean a "memory leak", but I've
> > > monitor
> > > > memory and it's very well... See perform.jog  - Is there another
> thing
> > > that
> > > > may be suggested ?
> > > > Another thing that I don't undestand of this graph is the "y" axis,
> is
> > it
> > > a
> > > > "percentage" of what? and if the meassure is in milliseconds.  Sorry,
> I
> > > > didn't find it in the user manual or app help.
> > > >
> > > > About the other graph, why I'm getting those "hops" or points ??
> > > >
> > > > Thanks a lot !!!
> > > >
> > > >
> > > > ---------------------------------------------------------------------
> > > > 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