You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by ankitR <ju...@yahoo.co.in> on 2010/08/23 08:20:06 UTC
Jmeter time vs User Percieved time
Hi,
I am using Jmeter to stress the server with 50 concurrent users.At the same
time I note down the manual time( i.e user perceived time as i am interested
in the high level UI rendering time) by opening a browser and following the
same steps as in Jmeter script. I am using three different machines for this
( app server, Jmeter scripts & for browser). when I note the time taken by
Jmeter for loading the jsp i am interested in it shows - average time = 19s
and 90% line = 35s, whereas the manual time for loading the same page is 7s.
I have used cache manager in Jmeter and also am using download all embedded
resources so that jmeter can emulate the browser behaviour to some extent.
since Jmeter is not a browser and doesn't give the UI rendering time, I
expected the manual time to be more than the jmeter time not vice-versa.
(the times mentioned earlier are averages of multiple runs).
Now I am thinking should I even compare the two times?
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-time-vs-User-Percieved-time-tp2644274p2644274.html
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: Jmeter time vs User Percieved time
Posted by Felix Frank <ff...@mpexnet.de>.
On 08/23/10 08:20, ankitR wrote:
>
> Hi,
>
> I am using Jmeter to stress the server with 50 concurrent users.At the same
> time I note down the manual time( i.e user perceived time as i am interested
> in the high level UI rendering time) by opening a browser and following the
> same steps as in Jmeter script. I am using three different machines for this
> ( app server, Jmeter scripts & for browser). when I note the time taken by
> Jmeter for loading the jsp i am interested in it shows - average time = 19s
> and 90% line = 35s, whereas the manual time for loading the same page is 7s.
> I have used cache manager in Jmeter and also am using download all embedded
> resources so that jmeter can emulate the browser behaviour to some extent.
> since Jmeter is not a browser and doesn't give the UI rendering time, I
> expected the manual time to be more than the jmeter time not vice-versa.
> (the times mentioned earlier are averages of multiple runs).
> Now I am thinking should I even compare the two times?
You absolutely should, but your base assumption may be wrong.
I remember a stress test where Jmeter consistently reported overlong
times. The reason where background scripts. The web developer managed
site statistics by embedding a 1x1 transparent gif in each page, the
source being an elaborate call to a php script. When the load was high,
the web servers took several seconds to process that script. However,
the page was displayed fine way before that. So while Jmeter patiently
waited 20 seconds or more for each page, they were visible much sooner.
Use the excellent FireBug extension for Firefox (specifically, its "Net"
tab) to monitor loading times for DOM content and all embedded elements
respectively. Google's Chrome has a similar feature built in.
Cheers,
Felix
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Jmeter time vs User Percieved time
Posted by Felix Frank <ff...@mpexnet.de>.
On 08/23/10 10:30, ankitR wrote:
>
> Hi,
>
> Thanks for the help. However I have some doubts. The manual time that I am
> taking is when the server is loaded with the 50 users spawned by Jmeter. So
> shouldn't the manual time also include the the time taken by all the
> background scripts running on the webserver?
> Secondly, I tried using firebug earlier and from what I understand it gives
> you the load time of all the dom content, js. images ets loaded with the
> page. I have enabled the download all embedded resources checkbox in my http
> requests so Jmeter also gives me that time. I am not clear as to how firebug
> will help me. would you please elaborate a bit. I am pretty new to jmeter
> and other test tools. Please excuse my limited knowledge.
> Thirdly, the manual time is taken when the server is loaded with infact 51
> users (50 jmeter + 1 manual). For just 1 jmeter user it shows 448ms as
> opposed to 35s when there are 50 users. What I want is what the end user
> sees when the server is having a load of 50 users, so I don't feel comparing
> with 1 jmeter user will help. could you elaborate a bit upon what you meant?
> If it helps I noted the time when there is only 1 manual user( Jmeter test
> is not started) and it is 6.5s
I believe that was a misunderstanding on Deepak's part. The time you
perceive during the Jmeter run is in fact what you want to compare to
Jmeter's measured time.
Use Firebug to find out how the total load times (including each last
embedded element) relates to the time you perceive as the actual load
time, i.e. how long it takes until you see the page.
There may be single images or scripts that stop Firefox and Jmeter from
considering a page as completely loaded, but the Browser may already
display what appears to be an actually complete page. It is important to
determine if that time is consistently below the machine measured time.
HTH,
Felix
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Jmeter time vs User Percieved time
Posted by gunjan <gu...@germinait.com>.
Hi
I recorded a script,compared its result with firebug,the time for the
actions given by firebug are comparatively more than that of jmeter,could
any one please suggest as in what steps should be followed to elminate this
mismatch.
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-time-vs-User-Percieved-time-tp2644274p4345915.html
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: Jmeter time vs User Percieved time
Posted by Felix Frank <ff...@mpexnet.de>.
If there is anything funny going on, Firebug should indicate that.
Compare Firebug's output with the timings as reported by the View Result
Tree Listener. Expand a sample to examine the timings of all sub-samples.
HTH,
Felix
On 08/24/10 10:45, ankitR wrote:
>
> Hi,
>
> I tried taking the manual time on the same machine from which jmeter scripts
> are running but still i am getting the user perceived time as much less than
> what jmeter shows ( 12s vs 26s).
> Can you suggest me any way in which i can find out whether the server is
> running any background scripts?
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Jmeter time vs User Percieved time
Posted by ankitR <ju...@yahoo.co.in>.
Hi,
I tried taking the manual time on the same machine from which jmeter scripts
are running but still i am getting the user perceived time as much less than
what jmeter shows ( 12s vs 26s).
Can you suggest me any way in which i can find out whether the server is
running any background scripts?
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-time-vs-User-Percieved-time-tp2644274p2650886.html
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: Jmeter time vs User Percieved time
Posted by Deepak Goel <de...@gmail.com>.
Hey
Namaskara~Nalama~Guten Tag
What you are then looking out for is the average response time of 50 users.
Deepak
--
Keigu
Deepak
+91-9765089593
deicool@gmail.com
http://www.simtree.net
Skype: thumsupdeicool
Google talk: deicool
Blog: http://loveandfearless.wordpress.com
Facebook: http://www.facebook.com/deicool
"Contribute to the world, environment and more : http://www.gridrepublic.org
"
On Mon, Aug 23, 2010 at 2:00 PM, ankitR <ju...@yahoo.co.in> wrote:
>
> Hi,
>
> Thanks for the help. However I have some doubts. The manual time that I am
> taking is when the server is loaded with the 50 users spawned by Jmeter.
> So
> shouldn't the manual time also include the the time taken by all the
> background scripts running on the webserver?
> Secondly, I tried using firebug earlier and from what I understand it gives
> you the load time of all the dom content, js. images ets loaded with the
> page. I have enabled the download all embedded resources checkbox in my
> http
> requests so Jmeter also gives me that time. I am not clear as to how
> firebug
> will help me. would you please elaborate a bit. I am pretty new to jmeter
> and other test tools. Please excuse my limited knowledge.
> Thirdly, the manual time is taken when the server is loaded with infact 51
> users (50 jmeter + 1 manual). For just 1 jmeter user it shows 448ms as
> opposed to 35s when there are 50 users. What I want is what the end user
> sees when the server is having a load of 50 users, so I don't feel
> comparing
> with 1 jmeter user will help. could you elaborate a bit upon what you
> meant?
> If it helps I noted the time when there is only 1 manual user( Jmeter test
> is not started) and it is 6.5s
>
> Thanks & Regards.
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Jmeter-time-vs-User-Percieved-time-tp2644274p2644364.html
> 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: Jmeter time vs User Percieved time
Posted by ankitR <ju...@yahoo.co.in>.
Hi,
Thanks for the help. However I have some doubts. The manual time that I am
taking is when the server is loaded with the 50 users spawned by Jmeter. So
shouldn't the manual time also include the the time taken by all the
background scripts running on the webserver?
Secondly, I tried using firebug earlier and from what I understand it gives
you the load time of all the dom content, js. images ets loaded with the
page. I have enabled the download all embedded resources checkbox in my http
requests so Jmeter also gives me that time. I am not clear as to how firebug
will help me. would you please elaborate a bit. I am pretty new to jmeter
and other test tools. Please excuse my limited knowledge.
Thirdly, the manual time is taken when the server is loaded with infact 51
users (50 jmeter + 1 manual). For just 1 jmeter user it shows 448ms as
opposed to 35s when there are 50 users. What I want is what the end user
sees when the server is having a load of 50 users, so I don't feel comparing
with 1 jmeter user will help. could you elaborate a bit upon what you meant?
If it helps I noted the time when there is only 1 manual user( Jmeter test
is not started) and it is 6.5s
Thanks & Regards.
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-time-vs-User-Percieved-time-tp2644274p2644364.html
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: Jmeter time vs User Percieved time
Posted by Deepak Goel <de...@gmail.com>.
Hey
Namaskara~Nalama~Guten Tag
The manual test is being run only with 1 user whereas Jmeter is being run
with 50 users. Have you tried running Jmeter with 1 user and then comparing
the times?
Deepak
--
Keigu
Deepak
+91-9765089593
deicool@gmail.com
http://www.simtree.net
Skype: thumsupdeicool
Google talk: deicool
Blog: http://loveandfearless.wordpress.com
Facebook: http://www.facebook.com/deicool
"Contribute to the world, environment and more : http://www.gridrepublic.org
"
On Mon, Aug 23, 2010 at 11:50 AM, ankitR <ju...@yahoo.co.in> wrote:
>
> Hi,
>
> I am using Jmeter to stress the server with 50 concurrent users.At the same
> time I note down the manual time( i.e user perceived time as i am
> interested
> in the high level UI rendering time) by opening a browser and following the
> same steps as in Jmeter script. I am using three different machines for
> this
> ( app server, Jmeter scripts & for browser). when I note the time taken by
> Jmeter for loading the jsp i am interested in it shows - average time = 19s
> and 90% line = 35s, whereas the manual time for loading the same page is
> 7s.
> I have used cache manager in Jmeter and also am using download all embedded
> resources so that jmeter can emulate the browser behaviour to some extent.
> since Jmeter is not a browser and doesn't give the UI rendering time, I
> expected the manual time to be more than the jmeter time not vice-versa.
> (the times mentioned earlier are averages of multiple runs).
> Now I am thinking should I even compare the two times?
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Jmeter-time-vs-User-Percieved-time-tp2644274p2644274.html
> 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
>
>