Pagespeed bug / misleading hint

26 views
Skip to first unread message

Andreas Steffan

unread,
Jul 26, 2016, 5:21:38 AM7/26/16
to pagespeed-insights-discuss
Hallo everyboby,

for https://www.contentreich.de/who.html, Pagespeed suggests "Consider Fixing: ... Only about 6% of the final above-the-fold content could be rendered with the full HTML response"

The page does not use any external at all.

Looking at the screenshot, I see the issue seems to be caused by the image (which is inline).

Alternatively, I tried an inline SVG image placeholder (which is commented out in  the code). Pagespeed shows up the same issue.

What's wrong here?

Thank you
Andreas

Rick Steinwand

unread,
Jul 26, 2016, 8:18:44 AM7/26/16
to pagespeed-insights-discuss
Too many bytes are needed to render the what you see. 

Are you inlining critical css and post loading the rest? https://jonassebastianohlsson.com/criticalpathcssgenerator/
Are you post loading non critical scripts? https://github.com/filamentgroup/loadCSS

It's possible there's nothing you can do.

Andreas Steffan

unread,
Jul 26, 2016, 8:45:54 AM7/26/16
to pagespeed-insights-discuss
The entire page weighs in at 76k, the inline png is 16k. The (commented) svg is 657 bytes, so using the SVG instead gets total weight down to 60k.

The result remains the same. (SVG version is at https://www.contentreich.de/bar.html)

All CSS is in the header and the page does not use anything external to the HTML document.

Are you sure 60k is still too much?

If that's the case, the message "Only about 6% of the final above-the-fold content could be rendered with the full HTML response" should probably be changed because it is wrong.

regards
Andreas

Rick Steinwand

unread,
Jul 26, 2016, 9:12:19 AM7/26/16
to pagespeed-insights-discuss
You have a huge amount of inline css and fonts in the head. Sorry, this not a bug.

Ideally you want the viewport to be 14.6 mb (compressed) or less (did you even read this): https://developers.google.com/speed/docs/insights/PrioritizeVisibleContent

FWIW, I got rid of a similar message by moving an image down a couple lines.

If you want to get rid of this message, try my recommendations.

Don't inline your entire css. It's not cacheable that way.
Inline only critical css and post-load the rest: https://jonassebastianohlsson.com/criticalpathcssgenerator/
Defer loading scripts and post-load non critical scripts: https://github.com/filamentgroup/loadCSS

This page is heavier and passes all but Google Analytics: http://www.intercepteft.com/achprocessing.html

Andreas Steffan

unread,
Jul 26, 2016, 9:32:05 AM7/26/16
to pagespeed-insights-discuss
Guess I missed the message about the size one page further down. :/

And I guess you are right about it.

Still, I find the message "could be rendered with the full HTML response" misleading - and wrong.

Thanks a lot Rick!
Andreas

Rick Steinwand

unread,
Jul 26, 2016, 9:56:40 AM7/26/16
to pagespeed-insights-discuss
It's a confusing message until you read the linked page.

Of course many of the rules change for HTTP/2, but it'll be awhile for that.

Andreas Steffan

unread,
Jul 26, 2016, 10:26:58 AM7/26/16
to pagespeed-insights-discuss
Finally, I can confirm you were indeed right and trimming the css in the head down to the minimum solved my issue.

Again, thanks Rick.

Rick Steinwand

unread,
Jul 26, 2016, 11:22:43 AM7/26/16
to pagespeed-insights-discuss
Thanks for the update.

Glad you got the error to go away.
Reply all
Reply to author
Forward
0 new messages