Re: [JMeter][RFC] JMeter report generator based on JAMon

Siegfried Goeschl <[email protected]>
Newsgroups gmane.comp.jakarta.cactus.devel
Organization IT20one GmbH
Message-ID <[email protected]>
Hi Sebastian,

see my comments below

Cheers,

Siegfried Goeschl

On 25.07.11 16:11, sebb wrote:
> On 25 July 2011 14:14, Siegfried Goeschl<[email protected]>  wrote:
>> Hi Sebastian,
>>
>> a few clarifications
>>
>> +) the JMeter report is consumed by a STAX parser so this is basically a
>> post-processor of the JTL file
> JMeter already contains code to analyse JT:L files - might be better
> to use that, as any changes to the format will be handled by JMeter.
>
since I don't know the JMeter code base where can I find the code to 
parser a JTL file?
>> +) the label and response time of the HTTP invocation is fed into JAMon
>> +) with a bit of manual code the underlying JAMon data model is written to a
>> HTML file
> So you are only using the JAMon data model, is that correct?
>
> JMeter also contains code to do summarising and various other
> statistical analyses for the GUI Listeners.

IMHO not good enough because I need the fulfill SLA's and require 
mapping response times into intervals as shown in the "Page Detail 
Table" section.

>> +) currently there is no way to customize the output
>>
>> I also played with CSV output but it is not significantly faster than XML in
>> my case but I will re-check.
> CSV output will be much smaller and is less expensive at run-time.
> It's also easier to parse - and again, JMeter has the code to parse the file.

Not sure about it but I think the XML output is the way to go because I 
dump request and response on errors ...

>> Cheers,
>>
>> Siegfried Goeschl
>>
>> On 25.07.11 14:25, sebb wrote:
>>> On 24 July 2011 15:19, Siegfried Goeschl<[email protected]>
>>>   wrote:
>>>> Hi folks,
>>>>
>>>> I'm using JMeter for some large-scale performance tests and had problems
>>>> with the current approach of using XSLT to generate the HTML report since
>>>> my
>>>> JTL files exceed two gigabytes. Therefore I implemented a new reporting
>>>> backend based on JAMon (see http://jamonapi.sourceforge.net/)
>>>>
>>>> Code : https://github.com/sgoeschl/jmeter-sla-report
>>>> Sample Report :
>>>> http://people.apache.org/~sgoeschl/download/jmeter/summary-report.html
>>>>
>>>> If this is of general interest I would like to contribute the code -
>>>> please
>>>> note that
>>>>
>>>> +) JAMon is under BSD licence
>>>> +) the latest JAMon libraries are currently not available at Maven
>>>> Central
>>>>
>>>> Feedback appreciated
>>> I'm a bit unclear on how JAMon is being used.
>>> It looks to me like JAMon is a monitoring system, rather than a
>>> reporting system, so how much of the JAMon code is really being used?
>>>
>>> Also, how easy is it to customise the output?
>>>
>>> ==
>>>
>>> By the way, for long-running tests, it's a lot more efficient to use
>>> CSV output instead of XML.
>>> CSV does not support all the possible data types that XML does, but
>>> the additional fields in XML are not normally needed for reports.
>>>
>>>> Siegfried Goeschl
>>>>
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [email protected]
>>>> For additional commands, e-mail: [email protected]
>>>>
>>>>
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.