Support » Plugin: Autoptimize » autoptimize not working with WPML in multiligual approach

  • Resolved wci

    (@wci)


    Dear support team,

    Our page loading speed differs while using different WPML language (EN faster than DE) approach.

    After further investigation even in help with the WPML support, we found out that the autoptimize plugin in use differs our page loading speed while using different language approach.

    Testing:
    I reactivated cache with all 93 plugins activated
    autoptimize and wp supercache with these results
    Screenshot > https://www.screencast.com/t/tfWc7x1q
    approx EN 3,74 DE 6,90 + 7,25 sec

    Then deactivated autoptimize plugin
    Screenshot https://www.screencast.com/t/fSutQWWT0h
    apporx EN 7,76 DE 7,76 + 7,31

    >>>THAT MEANS THAT THE “AUTOPTIMIZE PLUGIN” CACHE FOR WPML DE APPROACH IS NOT APPLICABLE IF THIS PLUGIN GOT ACTIVATED
    in other words autoptimize does the job for the EN content but for the DE content, it’s not working for whatever reason.

    Please help to find out why and how to solve this to be able to use both languages in same fast loading results by using your plugin.

    Best regards, Alex

Viewing 10 replies - 1 through 10 (of 10 total)
  • Plugin Author Optimizing Matters

    (@optimizingmatters)

    That’s weird ..

    I would need more info then just the total loading time to understand what is happening, do you have such a site where AO is currently active so I could investigate more in detail what is (not) happening Alex?

    Hi again,

    sure just use this our EN and DE page and check your results.

    For testing just use pingdom tools location in Frankfurt Germany for EN (loading in 3 sec) or DE (loading in 7,81 sec) approach.

    Autoptimize is activated as you can see in this screenshot > https://nimb.ws/iTlq9y

    Best regards, Alex

    Plugin Author Optimizing Matters

    (@optimizingmatters)

    I’m doing some tests, but see big variations and sometimes errors. Could it be you’re making changes at this moment wci ?

    Hi again,

    Yes I’ve done some changes.

    I’ve just realized that the server was running out of disk space and I deleted some backups to proceed
    AND
    I’ve deactivated the cloudflare DNS approach to point directly to the server not using the cache form their end to be able start with direct server related investigation.

    Please go ahead with your investigation.

    Best, Alex

    • This reply was modified 2 months, 2 weeks ago by wci.
    • This reply was modified 2 months, 2 weeks ago by wci.

    Hi again,
    FYI – I’ve had to make some theme related testings and deactivated wp super cache and autoptimize.

    Now I am finished with this and re-activated both plugins back.
    You can go now further with your investigation. I will not disturb you again today.

    Thanks, Alex

    Plugin Author Optimizing Matters

    (@optimizingmatters)

    OK, did 2 tests on webpagetest.org (which offers much more detailed and statistically more relevant tests) with location Germany, unlimited bandwidth, Chrome, doing 9 runs each test.

    -> results for EN: https://www.webpagetest.org/result/191108_47_7e1bfcdfb90efd01521761581cbf858b/
    -> results DE: https://www.webpagetest.org/result/191108_RT_27642ae4979acbff4fdae4536756db5b/
    -> and most interestingly comparison of all KPI for the 2*9 tests in graphs here

    Based on this I’m afraid (or happy, depends on how you look at it) that there is no significant difference in performance of the 2 versions.

    I also had a look at how Autoptimize was working on the 2 versions and see no difference, in both cases the CSS & JS & HTML are optimized properly.

    Hope this reassures you to some extent!

    Have a nice weekend!
    frank

    Hi Frank,

    thank you for your investigation. Unfortunately, I cannot confirm the same page loading time results you’ve delivered.

    -> results for mydomain.com EN: https://www.webpagetest.org/result/191109_Q5_f4902becf20015013668c824ac9100f2/
    Load time 5.778s
    -> results mydomain.de referring to mydomain.com/de DE: https://www.webpagetest.org/result/191109_E1_43640241d8f78fc7265057719dd27ccd/
    Load time 14.377s

    Do you have any further ideas about why my results differ by using the following cache services:

    @own server
    cache related
    – cpnginx for http2 approach
    – autoptimize (wp plugin)
    – wp super cache (wp plugin)

    @external services from cloudflare
    – cloudflare cache
    with internal speed related settings
    – auto minify (javascript, css, html)
    – Brotli (Speed up page load times for your visitor’s HTTPS traffic by applying Brotli compression)
    – Rocket Loader (Improve the paint time for pages which include JavaScript)

    + additional Cloudflare services in use at our own server installed
    — CF railgun (Accelerate delivery of dynamic content.)
    — wrangler to enable cloudflare workers (settings to be made within CF account)
    — Cloudflare Page Cache (wp plugin)

    I am looking forward to your helping information to reach the goal of worldwide page loadings in under 2 seconds.

    Best regards, Alex

    Plugin Author Optimizing Matters

    (@optimizingmatters)

    afraid I don’t have the time to continue investigating this in detail Alex, but;
    * from AO’s point of view all looks OK for both tests
    * the mere fact that you do the test for DE with mydomain.de referring to mydomain.com/de makes for the first of many small differences, as that referring is in fact a redirect (301) which in itself takes some time already
    * when comparing the individual run results of the 2 tests, you’ll see that for the EN one you have 14s, 12s & 6s (which was the median result based on speed index score, but that number is skewed as apperantly the autoptimized CSS was missing in action) which is not _that_ different from the 21s (page cache miss resulting to longer time to first byte), 14s and 13s
    * because of the previous point, I would do 9 instead of 3 testruns to have results that is statistically more reliable (median of 3 vs median of 9 tests is a big difference)

    and “under 2s” worldwide will be … hard. my 2c; try getting rid of some of the JS (over 1 MB), CSS (almost 0,5MB) and fonts (over half an MB).

    best of luck (and don’t forget to enjoy the weekend) 🙂
    frank

    Hi Frank,

    thanks again for those helpful information. Maybe you have also advice in –

    try getting rid of some of the JS (over 1 MB), CSS (almost 0,5MB) and fonts (over half an MB).

    Any ideas to proceed with that?

    Best regards, Alex

    P.S. please enjoy your weekend too!!!

    Plugin Author Optimizing Matters

    (@optimizingmatters)

    have a look at plugin organizer or gonzales 🙂

Viewing 10 replies - 1 through 10 (of 10 total)
  • You must be logged in to reply to this topic.