You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@jmeter.apache.org by bu...@apache.org on 2018/06/25 05:06:22 UTC
[Bug 62486] New: Report Statics Table showing incorrect 90th pct
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Bug ID: 62486
Summary: Report Statics Table showing incorrect 90th pct
Product: JMeter
Version: 4.0
Hardware: PC
Status: NEW
Severity: normal
Priority: P2
Component: Main
Assignee: issues@jmeter.apache.org
Reporter: philip.webb@iag.com.au
Target Milestone: ---
The 90th percentile being reported in the Statistics Table can be different to
the 90th percentile shown at the 90th percentile on the "Response Times
Percentiles" graph (which is be viewed by hovering with the mouse at the 90.0
line).
I have calculated the 90th percentile independently. The graph is correct, the
value on the Statistics table is wrong.
Appears to happen on larger samples (I noticed it on a label with sample count
of > 200,000 )
Just a guess - but may be due to higher value responses being 'cut off' within
an internal array during the calculation (I made the same mistake initially)..
I can attach a sample CSV results file, if required (I will need to slightly
doctor it for business reasons).
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Philip Webb <ph...@iag.com.au> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEEDINFO |NEW
--- Comment #6 from Philip Webb <ph...@iag.com.au> ---
We generate the report using the output from a "Summary Report" listener thread
(we actually load the Summary Report onto our Jenkins server so our users can
see our results)
So if the Summary Report Data is at D:\JMETER\txn90thResults.csv,
and I want the Jmeter html report to go to D:\html,
The command I run is:
jmeter -g D:\JMETER\txn90thResults.csv -o D:\htmlreport
Note that I have had to edit the lablels in the sample I gave you for business
privacy reasons.
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
--- Comment #7 from Philip Webb <ph...@iag.com.au> ---
oops, in my previous comment I meant to say I want the Jmeter html report to go
to D:\htmlreport (not D:\html)
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Philippe Mouawad <p....@ubik-ingenierie.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |NEEDINFO
--- Comment #8 from Philippe Mouawad <p....@ubik-ingenierie.com> ---
(In reply to Philip Webb from comment #7)
> oops, in my previous comment I meant to say I want the Jmeter html report to
> go to D:\htmlreport (not D:\html)
Hello
You can increase :
backend_metrics_window=100
backend_metrics_large_window=5000
which will increase precision of the percentile.
We will do related to this in:
- https://github.com/apache/jmeter/pull/380
Regards
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Felix Schumacher <fe...@internetallee.de> changed:
What |Removed |Added
----------------------------------------------------------------------------
OS| |All
Status|NEW |NEEDINFO
--- Comment #1 from Felix Schumacher <fe...@internetallee.de> ---
Please check, if #61130 and/or #61725 are the same reported bugs, as yours.
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Philip Webb <ph...@iag.com.au> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEEDINFO |NEW
--- Comment #3 from Philip Webb <ph...@iag.com.au> ---
sample has been attached - you should be able to replicate issue so changing
status top NEW
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
--- Comment #2 from Philip Webb <ph...@iag.com.au> ---
Created attachment 35992
--> https://bz.apache.org/bugzilla/attachment.cgi?id=35992&action=edit
csv file used to generate reprot
Attached a sample CSV file, which I use to generate a html report (using jmeter
-g command)
On the resulting Jmeter report:
The Stats Table gives a 90th of 3528.00 for "LOTS_OF_TXN"
The"Response Time Percentile" graph give a 90th of 3379 for "LOTS_OF_TXN"
My independent calculation is 3379. That is the graph is exactly right for
this file, the Stats Table is not.
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
--- Comment #4 from Philip Webb <ph...@iag.com.au> ---
Oh, should of mentioned I don't believe this a duplicate.
It is similar to https://bz.apache.org/bugzilla/show_bug.cgi?id=61130 .. but
has a major difference: The larger the dataset, my observation is that it seems
the larger the error in the stats table. That is, it's not just a matter of a
subtle rounding issue. Also, the median was slightly out in my results, but
the 90th percentile was far more affected (why I think it’s a real problem and
to do with the size of the sample)
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Philippe Mouawad <p....@ubik-ingenierie.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |WORKSFORME
Status|NEEDINFO |RESOLVED
--- Comment #9 from Philippe Mouawad <p....@ubik-ingenierie.com> ---
No feedback, closing
--
You are receiving this mail because:
You are the assignee for the bug.
[Bug 62486] Report Statics Table showing incorrect 90th pct
Posted by bu...@apache.org.
https://bz.apache.org/bugzilla/show_bug.cgi?id=62486
Philippe Mouawad <p....@ubik-ingenierie.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |NEEDINFO
CC| |p.mouawad@ubik-ingenierie.c
| |om
--- Comment #5 from Philippe Mouawad <p....@ubik-ingenierie.com> ---
(In reply to Philip Webb from comment #2)
> Created attachment 35992 [details]
> csv file used to generate reprot
>
> Attached a sample CSV file, which I use to generate a html report (using
> jmeter -g command)
>
> On the resulting Jmeter report:
>
> The Stats Table gives a 90th of 3528.00 for "LOTS_OF_TXN"
>
> The"Response Time Percentile" graph give a 90th of 3379 for "LOTS_OF_TXN"
>
> My independent calculation is 3379. That is the graph is exactly right for
> this file, the Stats Table is not.
Are you using the Web Report generated at end of test or on demand
OR are you using another element ?
Thank you for clarification.
--
You are receiving this mail because:
You are the assignee for the bug.