Re: how to submit/register a feature request for JMeter?

[Felix Schumacher]

Am 5. Oktober 2018 14:00:05 MESZ schrieb Mathijs Groen :
>Hi,
>
>What is the procedure/website for entering a feature request for
>JMeter?

Have you had a look at http://jmeter.apache.org/issues.html ? 

Is there anything you are missing on that page? 

Regards, 
 Felix

>
>Thanks,
>Mathijs Groen

-
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org

how to submit/register a feature request for JMeter?

[Mathijs Groen]

Hi,

What is the procedure/website for entering a feature request for JMeter?

Thanks,
Mathijs Groen

Re: Constant throughput timer feature request

[sebb]

On 10 May 2016 at 03:02, Oliver Erlewein  wrote:
> Hi,
>
> Thanks for that. Ill raise it in bugzilla as a CR.
>
> I tested it with -1 (and -100) and it looks like it executes each thread
> once only. Of course that is a 5min high level test so is by no means
> exhaustive. But I'm guessing nobody is using that anyway.

When I tried -1 it calculated the delay as negative so did not delay at all.
There must be something else in the plan that is causing a problem.

Note that using 0 will cause a very high delay and threads will
therefore appear to execute once only.

> As for the sufficiently high number, I am just unsure how much overhead it
> adds.

Exactly the same as using it normally.

But it should be possible to reduce that.

> Cheers Oliver
> On 9/05/2016 21:26, "sebb"  wrote:
>
> On 9 May 2016 at 03:54, Oliver Erlewein  wrote:
>> Hi,
>>
>> I have the need to control the throughput on the constant throughput timer
>> via a variable. That includes the possibility to switch it off. In a lot
> of
>> JMeter areas you can do that by passing a negative value (ex. -1). But not
>> here. -1 will do one iteration only.
>
> Huh?
>
> -1 is not treated specially by the code so why do you think it only
> affects a single iteration?
>
>> Can we change this behaviour? Clearly -1 is not sensible in a constant
>> throughput controller so why not use it to skip it's function?
>
> Feel free to raise a Bugzilla enhancement request.
>
> But a work-round is to use a sufficiently high value so JMeter won't
> need to insert pauses.
>
>> Cheers
>> Oliver
>
> -
> To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
> For additional commands, e-mail: user-h...@jmeter.apache.org

-
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org

Re: Constant throughput timer feature request

[Oliver Erlewein]

Hi,

Thanks for that. Ill raise it in bugzilla as a CR.

I tested it with -1 (and -100) and it looks like it executes each thread
once only. Of course that is a 5min high level test so is by no means
exhaustive. But I'm guessing nobody is using that anyway.

As for the sufficiently high number, I am just unsure how much overhead it
adds.

Cheers Oliver
On 9/05/2016 21:26, "sebb"  wrote:

On 9 May 2016 at 03:54, Oliver Erlewein  wrote:
> Hi,
>
> I have the need to control the throughput on the constant throughput timer
> via a variable. That includes the possibility to switch it off. In a lot
of
> JMeter areas you can do that by passing a negative value (ex. -1). But not
> here. -1 will do one iteration only.

Huh?

-1 is not treated specially by the code so why do you think it only
affects a single iteration?

> Can we change this behaviour? Clearly -1 is not sensible in a constant
> throughput controller so why not use it to skip it's function?

Feel free to raise a Bugzilla enhancement request.

But a work-round is to use a sufficiently high value so JMeter won't
need to insert pauses.

> Cheers
> Oliver

-
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org

Re: Constant throughput timer feature request

[sebb]

On 9 May 2016 at 03:54, Oliver Erlewein  wrote:
> Hi,
>
> I have the need to control the throughput on the constant throughput timer
> via a variable. That includes the possibility to switch it off. In a lot of
> JMeter areas you can do that by passing a negative value (ex. -1). But not
> here. -1 will do one iteration only.

Huh?

-1 is not treated specially by the code so why do you think it only
affects a single iteration?

> Can we change this behaviour? Clearly -1 is not sensible in a constant
> throughput controller so why not use it to skip it's function?

Feel free to raise a Bugzilla enhancement request.

But a work-round is to use a sufficiently high value so JMeter won't
need to insert pauses.

> Cheers
> Oliver

-
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org

Constant throughput timer feature request

[Oliver Erlewein]

Hi,

I have the need to control the throughput on the constant throughput timer
via a variable. That includes the possibility to switch it off. In a lot of
JMeter areas you can do that by passing a negative value (ex. -1). But not
here. -1 will do one iteration only.

Can we change this behaviour? Clearly -1 is not sensible in a constant
throughput controller so why not use it to skip it's function?

Cheers
Oliver

Re: JMeter feature request - Overriding timer

[Flavio Cysne]

You could try using multiple Thread Groups and a property to control
(inside a while controller, maybe) when these requests should be fire.

There's a plugin, from JMeter Plugins, that could help you synchronize
these requests.

There's also the Beanshell Sampler that you could use to implement your own
way of achieving it.
Em 06/11/2015 09:46, "Ričardas Šliapikas"  escreveu:

> Hello,
>
>
>
> I have the following situation. I have several HTTP requests in my test
> plan and some of the requests are „transations“, i.e. they must happen
> immediatelly one after another. Consider the following example:
>
>
>
>
>
> Quickview Printed summer dress is a „Transation controler“ and there are
> two requests which happen when I select „Printed Summer dress“ in webshop.
> So I want them to be executed without any timer. I could do this by puting
> a separate timer as a child of each request, but this is not very
> convenient if there are many requests. So I want to have one timer for most
> of my requests, but I want to be able to exclude some request sequesnces
> from this timer. Or to say in other words I want to override the „Gaussian
> Random Timer“ just for the two requests under „Quickview Printed summer
> dress“.
>
> This is a real situation in a webshop, because „Quickview Printed summer
> dress“ opens a new html page and then immediatelly submits ajax request to
> open popup.
>
>
>
> Maybe there is a way to do it? Or is it only possible with new feature?
>
>
>
>
>
>
>
>
>

JMeter feature request - Overriding timer

[Ričardas Šliapikas]

Hello,

I have the following situation. I have several HTTP requests in my test plan 
and some of the requests are "transations", i.e. they must happen immediatelly 
one after another. Consider the following example:

[cid:image001.png@01D118A1.EFC65990]

Quickview Printed summer dress is a "Transation controler" and there are two 
requests which happen when I select "Printed Summer dress" in webshop. So I 
want them to be executed without any timer. I could do this by puting a 
separate timer as a child of each request, but this is not very convenient if 
there are many requests. So I want to have one timer for most of my requests, 
but I want to be able to exclude some request sequesnces from this timer. Or to 
say in other words I want to override the "Gaussian Random Timer" just for the 
two requests under "Quickview Printed summer dress".
This is a real situation in a webshop, because "Quickview Printed summer dress" 
opens a new html page and then immediatelly submits ajax request to open popup.

Maybe there is a way to do it? Or is it only possible with new feature?

Re: how-to or feature request for loaded results into a listener

[Glenn Caccia]

It is primarily about standard Apache JMeter code.  For example, listeners 
Aggregrate Graph, Distribution Graph, Graph Results and Response Time 
Graph are all standard JMeter listeners.  They all have the ability to load 
results from a file using the BROWSE button.  However, if I have a collection 
of these listeners, I have to manually go to each listener and click the BROWSE 
button and load the exact same file.  It would be great if all these listeners 
could be directed to use the same external file for loading results.  Certainly 
it would be nice for 3rd party listeners to have the ability to participate in 
that feature also.




 From: sebb seb...@gmail.com
To: JMeter Users List user@jmeter.apache.org; Glenn Caccia 
gacac...@yahoo.com 
Sent: Wednesday, January 22, 2014 3:00 PM
Subject: Re: how-to or feature request for loaded results into a listener
 

On 22 January 2014 19:35, Glenn Caccia gacac...@yahoo.com wrote:
 I'm thinking of adding a feature request, but wanted to first make sure I'm 
 not missing something.  I currently use JMeter along with JMeter-Plugins and 
 use graphing listeners from both for analyzing results.  When the test is 
 actually running, these listeners are disabled, but then after the test I can 
 load the results from file to get the graphs.  I would like to compare 
 results from two different runs and was thinking of having a JMeter script 
 that only has the listeners (including JMeter-Plugins' Composite Graph), 
 duplicated twice for each run.  Once results were loaded from file, I could 
 easily compare the two runs and even create custom overlay views using the 
 Composite Graph.  The problem is that I have to manually click the BROWSE 
 button for each listener to load the exact same results file.  It works, but 
 it's a hassle.  It would be nice if all listeners could be triggered to 
 reload results from their indicated source files from a single
  toolbar button or something like that.

 Is there currently a way to achieve this?  Does this seem like a good feature 
 request?

Note that the JMeter Plugins are produced by a 3rd party.
You need to direct any feature requests that relate to their listeners
to the relevant developers.

This does not seem like it is anything to do with the standard Apache
JMeter code.


 Thanks,

 Glenn

-
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org

how-to or feature request for loaded results into a listener

[Glenn Caccia]

I'm thinking of adding a feature request, but wanted to first make sure I'm not 
missing something.  I currently use JMeter along with JMeter-Plugins and use 
graphing listeners from both for analyzing results.  When the test is actually 
running, these listeners are disabled, but then after the test I can load the 
results from file to get the graphs.  I would like to compare results from two 
different runs and was thinking of having a JMeter script that only has the 
listeners (including JMeter-Plugins' Composite Graph), duplicated twice for 
each run.  Once results were loaded from file, I could easily compare the two 
runs and even create custom overlay views using the Composite Graph.  The 
problem is that I have to manually click the BROWSE button for each listener to 
load the exact same results file.  It works, but it's a hassle.  It would be 
nice if all listeners could be triggered to reload results from their indicated 
source files from a single
 toolbar button or something like that.

Is there currently a way to achieve this?  Does this seem like a good feature 
request?

Thanks,

Glenn

Feature request...

[Robin D. Wilson]

I use the 'GUI' mode for most of my JMeter testing. I configure a 'Tree 
Listener' to show any errors, to help me debug - I check the
Log/Display Only: Errors box, so that it only shows errors encountered.

Is it possible to have the tree listener log the entire thread loop when it 
encounters an error? That is, let's say I have a
thread that ends up running through  10 samplers. Currently, when I encounter 
an error in sampler # 8, only that response is logged
in the Tree Listener. I would love to see samplers 1-8 logged when the error 
occurs in the 8th sampler of the loop - so that I can
trace the entire context of the failure through the test plan.

As it stands right now, if sampler #1 is 'login', and sampler #2 is show pageX, 
and sampler #3 is POST formY, ...; by the time I get
to sampler #8, the die may have already been cast - and I might not know why 
sampler #8 broke... But if I could see the entire
thread for that loop, I would be able to trace the steps back correctly. 
Basically, I want to retain the entire set of results for a
thread until the thread starts over at the beginning of the loop - then if the 
loop completes without exception, discard the result
set and start over again. If the loop throws an exception, log the retained 
results in the Tree Listener.


Also, while I'm at it. It would be nice if the Tree Listener could group 
results by thread. That is, create a folder for each
thread, so that the results of thread #1 can be reviewed independently from the 
results of thread #99 (this would be especially
useful when logging successful responses _and_ errors since it would keep the 
request/responses for a given thread in a group with
the same thread). As it stands right now - all threads record in groups for the 
sampler name for each instance of the request, so I
see all the results essentially grouped by sampled response by 'time' (I put 
'time' in scare quotes because it's not exactly by
'time', it depends on when the sampler gets the chance to write its results - 
which may or may not be sequenced by time as compared
to a different thread - I frequently see the results show up out of order based 
on their timestamps - but it is relatively close to
time ordered). I think this feature would be necessary in order for the above 
feature request to be useful - otherwise it would take
a lot of effort to identify which sampler shown in the response tree matched to 
each error.

--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com




-
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org