[Qgis-psc] (QGIS 2020 Grant Report) QGIS Server and performance monitoring

[Andreas Neumann]

Hi,

This is actually the funny thing. In the benchmarks that Marco B. did for
my employer, QGIS 3 was consistently slower in vector rendering, while in
Pauls tests it is the opposite.

Paul: can you clarify if you tested "sequentially" without parallel
requests? To me, if I interpret it correctly, this seems to be the case
here. Are you doing 20 subsequent, sequential, requests?

In the Solothurn benchmarks that Marco did, we almost always did parallel
requests - and we found out that the threaded rendering slows down
rendering if you have a lot of parallel requests, whereas, if you have no
or only 2-3 parallel requests, the threaded rendering is faster. In
general, in reality we often have 5-15 parallel requests in our
organization in Solothurn.

The other differences in Paul's and the Solothurn benchmarks is the size of
the requested images. In Paul's tests, the image sizes (width and height in
pixels) is usually quite small, but in reality we often have much larger
image requests (usually between 1000 and 2000 pixels in case of GetPrint
even higher. That is also one of the differences that need examination.

Greetings,
Andreas


On Fri, 13 Nov 2020 at 00:26, Nyall Dawson <nyall.dawson at gmail.com> wrote:

> On Fri, 13 Nov 2020 at 08:14, Tim Sutton <tim at kartoza.com> wrote:
> >
> > Hi Paul
> >
> > Thank you so much for the update and for doing this great work! Do you
> have any insights into the results? From what I can make out, 2.x releases
> are consistently a little faster than 3.x but not by much.
>
> This doesn't seem the case to me -- there's a number of tests where
> 3.x is slightly faster, and one where 3.x is magnitudes faster. Sure,
> there's some results where 2.x is faster, but it's hardly a conclusive
> trend!
>
> Nyall
>
> >
> > Regards
> >
> > Tim
> >
> > On Thu, Nov 12, 2020 at 10:17 AM Paul Blottiere <
> blottiere.paul at gmail.com> wrote:
> >>
> >> Dear PSC,
> >>
> >>
> >> Here is the report about the QEP
> >> https://github.com/qgis/QGIS-Enhancement-Proposals/issues/185.
> >>
> >> The whole QGIS-Server-PerfSuite has been upgraded to use 3.10 and 3.14
> releases
> >> side by side with 2.18 and master branch. Performances may be now
> monitored
> >> daily with the latest releases. Moreover, a simple anomalies detection
> mechanism
> >> has been implemented and a mail is sent if a regression is detected.
> >>
> >> Several scenarios have also been added to compare performance with the
> same
> >> data but relying on different providers (PostGIS, Spatialite,
> Geopackage and
> >> Shapefile).
> >>
> >> Finally, a simple mechanism based on multiprocessing has been
> implemented to
> >> simulate multi-clients situation. However, we don't have scenario based
> on this
> >> mechanism for now.
> >>
> >> For those interested, the last report is available here:
> >> http://test.qgis.org/perf_test/graffiti/2020_11_12_01_00/report.html.
> >>
> >> In spite of the limited scope of scenarios, there're very interesting
> results.
> >> Here is not the place to talk about it, but I'll communicate about it
> later.
> >>
> >>
> >> Regards.
> >> --
> >> Paul Blottiere
> >> QCooperative: https://www.qcooperative.net/
> >> Hytech Imaging: https://hytech-imaging.fr/
> >> _______________________________________________
> >> Qgis-psc mailing list
> >> Qgis-psc at lists.osgeo.org
> >> https://lists.osgeo.org/mailman/listinfo/qgis-psc
> >
> >
> >
> > --
> >
> ------------------------------------------------------------------------------------------
> >
> > Tim Sutton
> > Visit http://kartoza.com to find out about open source:
> >  * Desktop GIS programming services
> >  * Geospatial web development
> > * GIS Training
> > * Consulting Services
> > Tim is a member of the QGIS Project Steering Committee
> >
> -------------------------------------------------------------------------------------------
> > _______________________________________________
> > Qgis-psc mailing list
> > Qgis-psc at lists.osgeo.org
> > https://lists.osgeo.org/mailman/listinfo/qgis-psc
> _______________________________________________
> Qgis-psc mailing list
> Qgis-psc at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/qgis-psc



-- 

--
Andreas Neumann
QGIS.ORG board member (treasurer)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-psc/attachments/20201113/f2d6502b/attachment.html>