jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "PAULO LOURENCO (RE-ESI-Workflow)" <>
Subject RE: completely different time responses
Date Wed, 17 Dec 2008 11:08:21 GMT
Hello Adrian,

Thanks for the quick reply!
You can check my answers below.


-----Original Message-----
From: Fitzpatrick, Adrian [] 
Sent: terça-feira, 16 de Dezembro de 2008 20:14
To: JMeter Users List
Subject: RE: completely different time responses

Hi Paulo,

Is the test you are running quite heavy, both in terms of load applied to
your server  and (more to the point) load on the machine on which you are
running JMeter (in terms of CPU and/or JVM memory usage)?

> The answer to that one is simple, which is no. the test is compromised of a loop which
does what I mentioned just three times. Login, pick the first ticket, open the corresponding
form and submitting at the end. Each run takes about 4 seconds to complete and introduces
basically no load on any of the machines (server and client).

If so one theory about what might be happening is that the aggregate report
listener - quite a "heavy" component in terms of processing - is adding
additional load on the Jmeter machine, and as a result the machine is slowing
down. As a result JMeter can send less requests per second to server, and
therefore the load on the server side is reduced when the aggregate report
listener is enabled. This lower load on the server could then cause the
average response time to be better.

> This sounds like a very interesting theory! To be even more honest, taken into consideration
my knowledge of JMeter, this is the only theory I can cling to, since it is the only one that
makes sense.

This might be particularly likely to occur if JMeter and your test web server
are both running on the same machine

> This is not the case. My test machine is a almost brand new workstation which does nothing
else except testing the server, so load on this machine is close to zero.

This is just a theory - but might be worth looking into! Perhaps you can
compare hits/second as measured from the server side, for the two scenarios? 

> I will dig into this matter to see if this is indeed happening.

Hope this is some help.

> Once again, thanks for the prompt reply!




This message has been delivered to the Internet by the Revenue Internet e-mail service (OP)


To unsubscribe, e-mail:
For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message