Groups keyboard shortcuts have been updated
Dismiss
See shortcuts

Help with achieving Core Web Vitals target

83 views
Skip to first unread message

Domenico Porto

unread,
Dec 30, 2024, 10:37:33 AM12/30/24
to Chrome UX Report (Discussions)
Hi all.
It's a long time I'm struggling to achieve a positive result on Core Web Vitals on the wordpress site I manage (www.laviadellavita.it) but without any success.
I'm using WpRocket for caching and optimization (and for this I dismissed Jetpack plugin in order to avoid any correlation between them) but things are going worse, specially on mobile but on desktop too, instead of improving. 
The very useful cruxvis I just discovered confirmed this trend in these last weeks.
The TTFB is very high and consequentially LCP and FCP are high too.

I optimized the database and what it's loaded at the beginning, I moved almost all my images on webp format (instead of jpeg), I used the plugin " Disable Used CSS Fonts Preload" suggested by WpRocket team to avoid an heavy load of the fonts on page loading,  but without success.

However I can see that:
- the PageSpeedInsight display very good scores (around 100) for all the pages of the site
- the Gtmetrix results are good
- the Pingdom results are good
Only the test done with the site www.webpagetest.org presents sometimes results not very good but I'm not able to understand what I can do to reduce the first step due to the sum of the DNS, Connect, SSL and HTML (if this is a possible issue).

Can anybody help me to adress my effort to the root cause of the problem in order to finally see all green on Core Web Vitals?
Thanks in advance for your suggestion.
Regards,
Domenico

❄ Johannes Henkel

unread,
Dec 30, 2024, 2:41:36 PM12/30/24
to Domenico Porto, Chrome UX Report (Discussions)
I think paying attention to the TTFB is a good idea in this case, as it does seem quite high. When looking at the PHONE data for the origin, it looks like TTFB and LCP used to be great, move in a very similar way, and both started deteriorating in October 2024.
image.png


In the response headers for the site, I noticed that max-age for the Cache-control is 0, so I think this means that the browser will not cache the pages, to make sure that site updates are reflected immediately. Not necessarily unreasonable, but I'm also seeing X-Aruba-cache: MISS consistently (homepage and one other page on the site that I looked at) - so maybe this is a server side cache by your hosting company that's designed to help with TTFB, and it could be that the Cache-control 0 setting from WpRocket is interfering with that?

image.png

Here is the WpRocket documentation on this: https://docs.wp-rocket.me/article/80-browser-caching#html-and-max-age (especially see "HTML and max-age"). Perhaps setting a modest max age (an hour or whatever is acceptable for your site) and looking at the X-Aruba-Cache header to see whether it then indicates something other than a MISS would be a starting point?

The other thought I have is that if the site traffic changed significantly, e.g. due to an ad campaign, that may also explain the deterioration.
TTFB and LCP are measured from the start of the navigation (e.g. the user's click), so if there are slow redirects, they will also affect the LCP.

Please take these thoughts with a grain of salt - I don't have direct experience with WpRocket.
Good wishes!

--
You received this message because you are subscribed to the Google Groups "Chrome UX Report (Discussions)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chrome-ux-repo...@chromium.org.
To view this discussion visit https://groups.google.com/a/chromium.org/d/msgid/chrome-ux-report/5dcffb82-54fe-4851-a709-ead584dad7c9n%40chromium.org.

Domenico Porto

unread,
Dec 30, 2024, 5:39:15 PM12/30/24
to Chrome UX Report (Discussions), joha...@google.com, Chrome UX Report (Discussions), Domenico Porto
Hi Johannes,
really thanks for your suggestion.
I've installed the plugin suggested by WpRocket in the article you mentioned hoping it could manage the x-aruba cache MISS issue.
Anyway, how can I reproduce the test you indicated in the following screenshot so I can check by myself if the max-age=0 have been reset?

image.png

Of course I'm going to flush both caches, Aruba (that I can't remove because is embedded with the hosting and the plugin is automatically reactivated if I delete it) and WpRocket (that I could remove if needed).

In the next days I'll monitor the CWV changes and I'll let you know.

Regards,
Domenico

❄ Johannes Henkel

unread,
Dec 30, 2024, 6:21:09 PM12/30/24
to Domenico Porto, Chrome UX Report (Discussions)
Ah, I should have mentioned, my screenshot was from the Network panel in Chrome DevTools.
In Chrome, look for the three dots in the upper right corner, select "More Tools" -> "Developer Tools", then find the "Network" tab. You may need to reload the page before you see requests there. Then, find the main HTML resource (e.g. look for "www.laviadellavita.it" under "Name" for your homepage), and when you click on that, you can see the request and the response headers under Headers.

image.png

I'm now seeing   X-Aruba-Cache: HIT and   Cache-Control: max-age=2592000.
So, I think your change did something, it's using the cache, so if that cache works as advertised, it may help with TTFB.

However, I think max-age 2592000 is very high! I think this is measured in seconds, so it's 30 days (I did 2592000.0 / (60 * 60 * 24)).
I think this particular setting applies both to Chrome and the server side cache.
Perhaps it would be better to set it to something like 3600 (1 hour), so that users will see a new version of the pages as well, within a reasonable time.

Unfortunately I don't know what the best practice is for this - maybe someone else here knows?

Documentation of these headers is here:
Reply all
Reply to author
Forward
0 new messages