Recent LCP & FCP Issues with Web Stores only

  • Resolved Brian

    (@bricam)


    6 months, 3 weeks ago

    All Web Stories on my site have recently started causing LCP/ FCP issues. Upon investigating it looks like the issues are caused by render blocking resources loaded from the CDNs used by the plugin:

    • cdn.ampproject.org
    • fonts.googleapis.com

    And because Web Stories are AMP pages, WP Rocket doesn’t work to optimize these resources: “WP Rocket will deactivate these incompatible features: Minification, DNS Prefetch, LazyLoad, Optimize CSS delivery (Load CSS Asynchronously and Remove Unused CSS), Load JavaScript deferred, Delay JavaScript Execution, Google Fonts Optimization”

    Any idea why the js and css resources loaded from these CDNs have recently started causing delays in load time? Isn’t this supposed to be optimized out of the box? Is this a bug?

    The page I need help with: [log in to see the link]

Viewing 4 replies - 1 through 4 (of 4 total)
  • Plugin Author Pascal Birchler

    (@swissspidy)

    Wenn I run this story through PageSpeed Insights, the main LCP culprit I see is the slow time to first byte (TTFB). This is something you can address by enabling some server-side caching in WP-Rocket. That should help a lot already. I don’t see any warnings in PageSpeed about the AMP CDN. I recommend looking into such caching and then observe results again.

    Thread Starter Brian

    (@bricam)

    Hey Pascal, thanks for the quick response. I beg to differ, if you run the page through a tool like webpagetest.org and look at the waterfall graph, you’ll see the render blocking resources are those delivered by the CDNs I already mentioned. Furthermore, looking at PageSpeed Insights, Load Delay makes up the largest % of LCP – 44% and 1,700 ms at last check, whereas TTFB was only 16% of LCP at .630 ms

    As far as I can tell, it still looks like the resources loaded by the CDNs are the culprits. Am I missing something?

    Plugin Author Pascal Birchler

    (@swissspidy)

    In my quick testing using PageSpeed Insights, I saw that real users are experiencing TTFB of 1.4s and LCP of 2.0s on mobile. Since TTFB is relatively high and LCP includes TTFB, I made that suggestion.

    The waterfall you shared also shows a 1s document load time until the first bits of JS can be loaded.

    The rest of the waterfall is not unusual for an AMP document. Scripts like v0.mjs and amp-story-1.0.mjs must be loaded for a story, after that the rest can be fetched.

    That said, if you have a particular suggestion for improving the way AMP documents are structured, I recommend filing an issue at https://github.com/ampproject/amphtml. The Web Stories WordPress plugin just outputs Web Stories in the format that is expected, we don’t have much influence there. The AMP repo is where it’s at.

    Plugin Support Luckyna San

    (@luckynasan)

    Since there hasn’t been any further responses this topic will be marked as resolved. Please feel free to open a?new support topic?if you have any other questions. Thank you!

Viewing 4 replies - 1 through 4 (of 4 total)
  • You must be logged in to reply to this topic.

Tags