You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by rmiller <ro...@cobancorporation.com> on 2006/12/14 02:22:25 UTC
Throughput in Aggregate report
We use Pages Downloaded per Second as a performance metric in LoadRunner. In
side by side tests with LoadRunner and JMeter there dosen't seem to be any
correlation between that metric in LoadRunner and Throughput in JMeter. Can
someone please explain how to get the eaquivalent metric in JMeter?
Regards,
Ron
--
View this message in context: http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a7865207
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
Re: Any way to get more response time breakdowns using Jmeter?
Posted by sebb <se...@gmail.com>.
On 15/12/06, Christensen, Alan <al...@hp.com> wrote:
> Thanks for the response!
>
> This is an example where the documentation doesn't give enough
> information to avoid questions being asked on this forum. I don't mind
> asking the questions as the responses usually come back quick! But I
> bet half of my questions could be answered if there was a bit more
> detail in the documentation. For example, the parameters that can be
> configured in the listeners. It isn't obvious what all of these are,
> and how they are measured. For example, what does it mean to "save Sub
> Results" or "save field names"? I haven't found any place where these
> are defined. Hence I am left to guess, or to try to figure them out
> from the results that get logged, or ask a question. I would think that
> it would save a lot of time for you if these things simply got
> documented in either "help" or on the webpages. Or maybe I am just not
> looking in the right places?
>
Documentation patches and/or bug reports are always welcome via Bugzilla.
> BTW, what are "save field names" and "save sub results"?
Field names = Column names
Sub results are samples that were automatically generated - e.g. when
the HTTP sampler downloads embedded content. They are shown as
children of the main sample when using the View Tree Listener.
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
RE: Any way to get more response time breakdowns using Jmeter?
Posted by "Christensen, Alan" <al...@hp.com>.
Thanks for the response!
This is an example where the documentation doesn't give enough
information to avoid questions being asked on this forum. I don't mind
asking the questions as the responses usually come back quick! But I
bet half of my questions could be answered if there was a bit more
detail in the documentation. For example, the parameters that can be
configured in the listeners. It isn't obvious what all of these are,
and how they are measured. For example, what does it mean to "save Sub
Results" or "save field names"? I haven't found any place where these
are defined. Hence I am left to guess, or to try to figure them out
from the results that get logged, or ask a question. I would think that
it would save a lot of time for you if these things simply got
documented in either "help" or on the webpages. Or maybe I am just not
looking in the right places?
BTW, what are "save field names" and "save sub results"?
-----Original Message-----
From: sebb [mailto:sebbaz@gmail.com]
Sent: Friday, December 15, 2006 6:59 AM
To: JMeter Users List
Subject: Re: Any way to get more response time breakdowns using Jmeter?
On 14/12/06, Christensen, Alan <al...@hp.com> wrote:
> Most of the commercial testing tools provide a lot more information
> with respect to response time. For example, most of them report the
> response time to the first byte as well as the response time to the
last byte.
> Breakdowns of DNS time and content downloads are also typically
> included. Does Jmeter have any of these capabilities, and if so, how
> would I turn them on?
See:
http://jakarta.apache.org/jmeter/usermanual/listeners.html
some listeners support latency (time to first response).
Other timings are not currently collected - please feel free to create a
Bugzilla enhancement request with full details of what you are seeking.
Patches also welcome.
S.
---------------------------------------------------------------------
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
Re: Any way to get more response time breakdowns using Jmeter?
Posted by sebb <se...@gmail.com>.
On 14/12/06, Christensen, Alan <al...@hp.com> wrote:
> Most of the commercial testing tools provide a lot more information with
> respect to response time. For example, most of them report the response
> time to the first byte as well as the response time to the last byte.
> Breakdowns of DNS time and content downloads are also typically
> included. Does Jmeter have any of these capabilities, and if so, how
> would I turn them on?
See:
http://jakarta.apache.org/jmeter/usermanual/listeners.html
some listeners support latency (time to first response).
Other timings are not currently collected - please feel free to create
a Bugzilla enhancement request with full details of what you are
seeking. Patches also welcome.
S.
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Throughput in Aggregate report
Posted by git <gi...@cubicalland.com>.
That looks like a really interesting article - I shall read it all!
I also have written up my experiences of throughput testing with JMeter;
these can be found at www.deployview.com and nerds-central.blogspot.com.
Cheers
AJ
On Thu, 2006-12-14 at 18:45 -0500, Peter Lin wrote:
> http://jakarta.apache.org/tomcat/articles/performance.pdf
> http://cvs.apache.org/~woolfel/article.zip
>
>
> On 12/14/06, Ron Miller <ro...@cobancorporation.com> wrote:
> >
> > Peter, I searched for your article but couldn't find it. Would you
> > please send a link to it?
> >
> > Ron
> >
> > -----Original Message-----
> > From: Peter Lin [mailto:woolfel@gmail.com]
> > Sent: Thursday, December 14, 2006 3:24 PM
> > To: JMeter Users List
> > Subject: Re: Throughput in Aggregate report
> >
> > usually I look at the capacity of the network and determine if the
> > bytes/sec
> > is going to saturate the bandwidth or not. I have an article in the
> > wiki
> > that explains this stuff in detail.
> >
> > peter
> >
> > On 12/14/06, Jian Tong <To...@toronto.cbc.ca> wrote:
> > >
> > > Hi Peter,
> > >
> > > As you said bytes/sec and requests/sec are better metrics for load
> > > testing, then I hope to know what the benchmark is for bytes/sec and
> > > requests/sec from your experience ?
> > >
> > > Thanks
> > >
> > > Jian Tong
> > >
> > > >>> "Peter Lin" <wo...@gmail.com> 12/13/06 8:29 PM >>>
> > > jmeter will measure the requests per second. that isn't the same as
> > pages
> > > per second for a couple of reasons.
> > >
> > > 1. a page may have multiple images and stuff embedded
> > > 2. images are cached by the browser the first time it's loaded
> > > 3. not every page is the same, so page per second a poor measurement
> > of
> > > load
> > > 4. bytes per second or requests per second is a better measurement
> > > 5. what happens when the pages change and a page ends up having more
> > > images?
> > >
> > > loadRunner is a good product. Pages downloaded per second isn't all
> > that
> > > useful from a capacity and planning perspective. It is much better to
> > > measure bytes/second and requests/second.
> > >
> > > my bias opinion.
> > >
> > > peter
> > >
> > > On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
> > > >
> > > >
> > > > We use Pages Downloaded per Second as a performance metric in
> > > LoadRunner.
> > > > In
> > > > side by side tests with LoadRunner and JMeter there dosen't seem to
> > be
> > > any
> > > > correlation between that metric in LoadRunner and Throughput in
> > JMeter.
> > > > Can
> > > > someone please explain how to get the eaquivalent metric in JMeter?
> > > >
> > > > Regards,
> > > > Ron
> > > > --
> > > > View this message in context:
> > > >
> > >
> > http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a786
> > 5207
> > > > 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
> > > >
> > > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
> >
> >
--
git <gi...@cubicalland.com>
www.cubicalland.com
Re: Throughput in Aggregate report
Posted by Peter Lin <wo...@gmail.com>.
http://jakarta.apache.org/tomcat/articles/performance.pdf
http://cvs.apache.org/~woolfel/article.zip
On 12/14/06, Ron Miller <ro...@cobancorporation.com> wrote:
>
> Peter, I searched for your article but couldn't find it. Would you
> please send a link to it?
>
> Ron
>
> -----Original Message-----
> From: Peter Lin [mailto:woolfel@gmail.com]
> Sent: Thursday, December 14, 2006 3:24 PM
> To: JMeter Users List
> Subject: Re: Throughput in Aggregate report
>
> usually I look at the capacity of the network and determine if the
> bytes/sec
> is going to saturate the bandwidth or not. I have an article in the
> wiki
> that explains this stuff in detail.
>
> peter
>
> On 12/14/06, Jian Tong <To...@toronto.cbc.ca> wrote:
> >
> > Hi Peter,
> >
> > As you said bytes/sec and requests/sec are better metrics for load
> > testing, then I hope to know what the benchmark is for bytes/sec and
> > requests/sec from your experience ?
> >
> > Thanks
> >
> > Jian Tong
> >
> > >>> "Peter Lin" <wo...@gmail.com> 12/13/06 8:29 PM >>>
> > jmeter will measure the requests per second. that isn't the same as
> pages
> > per second for a couple of reasons.
> >
> > 1. a page may have multiple images and stuff embedded
> > 2. images are cached by the browser the first time it's loaded
> > 3. not every page is the same, so page per second a poor measurement
> of
> > load
> > 4. bytes per second or requests per second is a better measurement
> > 5. what happens when the pages change and a page ends up having more
> > images?
> >
> > loadRunner is a good product. Pages downloaded per second isn't all
> that
> > useful from a capacity and planning perspective. It is much better to
> > measure bytes/second and requests/second.
> >
> > my bias opinion.
> >
> > peter
> >
> > On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
> > >
> > >
> > > We use Pages Downloaded per Second as a performance metric in
> > LoadRunner.
> > > In
> > > side by side tests with LoadRunner and JMeter there dosen't seem to
> be
> > any
> > > correlation between that metric in LoadRunner and Throughput in
> JMeter.
> > > Can
> > > someone please explain how to get the eaquivalent metric in JMeter?
> > >
> > > Regards,
> > > Ron
> > > --
> > > View this message in context:
> > >
> >
> http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a786
> 5207
> > > 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
> > >
> > >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>
RE: Throughput in Aggregate report
Posted by Ron Miller <ro...@cobancorporation.com>.
Peter, I searched for your article but couldn't find it. Would you
please send a link to it?
Ron
-----Original Message-----
From: Peter Lin [mailto:woolfel@gmail.com]
Sent: Thursday, December 14, 2006 3:24 PM
To: JMeter Users List
Subject: Re: Throughput in Aggregate report
usually I look at the capacity of the network and determine if the
bytes/sec
is going to saturate the bandwidth or not. I have an article in the
wiki
that explains this stuff in detail.
peter
On 12/14/06, Jian Tong <To...@toronto.cbc.ca> wrote:
>
> Hi Peter,
>
> As you said bytes/sec and requests/sec are better metrics for load
> testing, then I hope to know what the benchmark is for bytes/sec and
> requests/sec from your experience ?
>
> Thanks
>
> Jian Tong
>
> >>> "Peter Lin" <wo...@gmail.com> 12/13/06 8:29 PM >>>
> jmeter will measure the requests per second. that isn't the same as
pages
> per second for a couple of reasons.
>
> 1. a page may have multiple images and stuff embedded
> 2. images are cached by the browser the first time it's loaded
> 3. not every page is the same, so page per second a poor measurement
of
> load
> 4. bytes per second or requests per second is a better measurement
> 5. what happens when the pages change and a page ends up having more
> images?
>
> loadRunner is a good product. Pages downloaded per second isn't all
that
> useful from a capacity and planning perspective. It is much better to
> measure bytes/second and requests/second.
>
> my bias opinion.
>
> peter
>
> On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
> >
> >
> > We use Pages Downloaded per Second as a performance metric in
> LoadRunner.
> > In
> > side by side tests with LoadRunner and JMeter there dosen't seem to
be
> any
> > correlation between that metric in LoadRunner and Throughput in
JMeter.
> > Can
> > someone please explain how to get the eaquivalent metric in JMeter?
> >
> > Regards,
> > Ron
> > --
> > View this message in context:
> >
>
http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a786
5207
> > 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
> >
> >
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Throughput in Aggregate report
Posted by Peter Lin <wo...@gmail.com>.
usually I look at the capacity of the network and determine if the bytes/sec
is going to saturate the bandwidth or not. I have an article in the wiki
that explains this stuff in detail.
peter
On 12/14/06, Jian Tong <To...@toronto.cbc.ca> wrote:
>
> Hi Peter,
>
> As you said bytes/sec and requests/sec are better metrics for load
> testing, then I hope to know what the benchmark is for bytes/sec and
> requests/sec from your experience ?
>
> Thanks
>
> Jian Tong
>
> >>> "Peter Lin" <wo...@gmail.com> 12/13/06 8:29 PM >>>
> jmeter will measure the requests per second. that isn't the same as pages
> per second for a couple of reasons.
>
> 1. a page may have multiple images and stuff embedded
> 2. images are cached by the browser the first time it's loaded
> 3. not every page is the same, so page per second a poor measurement of
> load
> 4. bytes per second or requests per second is a better measurement
> 5. what happens when the pages change and a page ends up having more
> images?
>
> loadRunner is a good product. Pages downloaded per second isn't all that
> useful from a capacity and planning perspective. It is much better to
> measure bytes/second and requests/second.
>
> my bias opinion.
>
> peter
>
> On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
> >
> >
> > We use Pages Downloaded per Second as a performance metric in
> LoadRunner.
> > In
> > side by side tests with LoadRunner and JMeter there dosen't seem to be
> any
> > correlation between that metric in LoadRunner and Throughput in JMeter.
> > Can
> > someone please explain how to get the eaquivalent metric in JMeter?
> >
> > Regards,
> > Ron
> > --
> > View this message in context:
> >
> http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a7865207
> > 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
> >
> >
>
Any way to get more response time breakdowns using Jmeter?
Posted by "Christensen, Alan" <al...@hp.com>.
Most of the commercial testing tools provide a lot more information with
respect to response time. For example, most of them report the response
time to the first byte as well as the response time to the last byte.
Breakdowns of DNS time and content downloads are also typically
included. Does Jmeter have any of these capabilities, and if so, how
would I turn them on?
Re: Throughput in Aggregate report
Posted by Jian Tong <To...@Toronto.CBC.CA>.
Hi Peter,
As you said bytes/sec and requests/sec are better metrics for load testing, then I hope to know what the benchmark is for bytes/sec and requests/sec from your experience ?
Thanks
Jian Tong
>>> "Peter Lin" <wo...@gmail.com> 12/13/06 8:29 PM >>>
jmeter will measure the requests per second. that isn't the same as pages
per second for a couple of reasons.
1. a page may have multiple images and stuff embedded
2. images are cached by the browser the first time it's loaded
3. not every page is the same, so page per second a poor measurement of load
4. bytes per second or requests per second is a better measurement
5. what happens when the pages change and a page ends up having more images?
loadRunner is a good product. Pages downloaded per second isn't all that
useful from a capacity and planning perspective. It is much better to
measure bytes/second and requests/second.
my bias opinion.
peter
On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
>
>
> We use Pages Downloaded per Second as a performance metric in LoadRunner.
> In
> side by side tests with LoadRunner and JMeter there dosen't seem to be any
> correlation between that metric in LoadRunner and Throughput in JMeter.
> Can
> someone please explain how to get the eaquivalent metric in JMeter?
>
> Regards,
> Ron
> --
> View this message in context:
> http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a7865207
> 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
>
>
Re: Throughput in Aggregate report
Posted by Peter Lin <wo...@gmail.com>.
sadly, jmeter does not provide that functionality :(
luckily for me, the times I've had to give performance numbers for capacity
planning and performance specs, it was in terms of bandwidth consumed and
requests/second.
peter
On 12/13/06, Ron Miller <ro...@cobancorporation.com> wrote:
>
> I agree with your argument about pages downloaded per second Peter; in
> fact this graph is off by default in LoadRunner. But my company has used
> this as a metric (among others) for years and is not about to change.
> That's why I'm trying to find a JMeter equivalent.
>
> -----Original Message-----
> From: Peter Lin [mailto:woolfel@gmail.com]
> Sent: Wednesday, December 13, 2006 5:30 PM
> To: JMeter Users List
> Subject: Re: Throughput in Aggregate report
>
> jmeter will measure the requests per second. that isn't the same as
> pages
> per second for a couple of reasons.
>
> 1. a page may have multiple images and stuff embedded
> 2. images are cached by the browser the first time it's loaded
> 3. not every page is the same, so page per second a poor measurement of
> load
> 4. bytes per second or requests per second is a better measurement
> 5. what happens when the pages change and a page ends up having more
> images?
>
> loadRunner is a good product. Pages downloaded per second isn't all that
> useful from a capacity and planning perspective. It is much better to
> measure bytes/second and requests/second.
>
> my bias opinion.
>
> peter
>
> On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
> >
> >
> > We use Pages Downloaded per Second as a performance metric in
> LoadRunner.
> > In
> > side by side tests with LoadRunner and JMeter there dosen't seem to be
> any
> > correlation between that metric in LoadRunner and Throughput in
> JMeter.
> > Can
> > someone please explain how to get the eaquivalent metric in JMeter?
> >
> > Regards,
> > Ron
> > --
> > View this message in context:
> >
> http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a786
> 5207
> > 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
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>
RE: Throughput in Aggregate report
Posted by Ron Miller <ro...@cobancorporation.com>.
I agree with your argument about pages downloaded per second Peter; in
fact this graph is off by default in LoadRunner. But my company has used
this as a metric (among others) for years and is not about to change.
That's why I'm trying to find a JMeter equivalent.
-----Original Message-----
From: Peter Lin [mailto:woolfel@gmail.com]
Sent: Wednesday, December 13, 2006 5:30 PM
To: JMeter Users List
Subject: Re: Throughput in Aggregate report
jmeter will measure the requests per second. that isn't the same as
pages
per second for a couple of reasons.
1. a page may have multiple images and stuff embedded
2. images are cached by the browser the first time it's loaded
3. not every page is the same, so page per second a poor measurement of
load
4. bytes per second or requests per second is a better measurement
5. what happens when the pages change and a page ends up having more
images?
loadRunner is a good product. Pages downloaded per second isn't all that
useful from a capacity and planning perspective. It is much better to
measure bytes/second and requests/second.
my bias opinion.
peter
On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
>
>
> We use Pages Downloaded per Second as a performance metric in
LoadRunner.
> In
> side by side tests with LoadRunner and JMeter there dosen't seem to be
any
> correlation between that metric in LoadRunner and Throughput in
JMeter.
> Can
> someone please explain how to get the eaquivalent metric in JMeter?
>
> Regards,
> Ron
> --
> View this message in context:
>
http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a786
5207
> 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
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Throughput in Aggregate report
Posted by Peter Lin <wo...@gmail.com>.
jmeter will measure the requests per second. that isn't the same as pages
per second for a couple of reasons.
1. a page may have multiple images and stuff embedded
2. images are cached by the browser the first time it's loaded
3. not every page is the same, so page per second a poor measurement of load
4. bytes per second or requests per second is a better measurement
5. what happens when the pages change and a page ends up having more images?
loadRunner is a good product. Pages downloaded per second isn't all that
useful from a capacity and planning perspective. It is much better to
measure bytes/second and requests/second.
my bias opinion.
peter
On 12/13/06, rmiller <ro...@cobancorporation.com> wrote:
>
>
> We use Pages Downloaded per Second as a performance metric in LoadRunner.
> In
> side by side tests with LoadRunner and JMeter there dosen't seem to be any
> correlation between that metric in LoadRunner and Throughput in JMeter.
> Can
> someone please explain how to get the eaquivalent metric in JMeter?
>
> Regards,
> Ron
> --
> View this message in context:
> http://www.nabble.com/Throughput-in-Aggregate-report-tf2817925.html#a7865207
> 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
>
>