Wrong LCP returned (compared to Chrome performance devtools + grafana)

[YannDuv]

FAQ

• Yes, my issue is not about variability or throttling.
• Yes, my issue is not about a specific accessibility audit (file with axe-core instead).

URL

https://www.se.com/ww/en/product/LC1D09M7/tesys-deca-contactor---3p(3-no)---ac-3---%3C%3D-440-v-9-a---220-v-ac-coil/?image=opti

What happened?

• Results coming from performance tab gives an average of 3.2s for LCP. I set throttling as 1.6Mbit + 150ms ; X4 CPU. Grafana gives similar results
• Results from lighthouse is somewhere between 5.5s and 7s on average ( lighthouse-metrics.com results )
• webPageTest are even better with 2.3s

What did you expect?

I was hoping to have something more consistent, at least with Chrome devtools. Something is off with lighthouse results for this page. If the issue is from the page I do not know where to start.
It would be helpful if Lighthouse explicitly names the Largest Content (as we can have in performance tab).

What have you tried?

No response

How were you running Lighthouse?

PageSpeed Insights, Chrome DevTools

Lighthouse Version

8.4.0 and 9.0.0

Chrome Version

96.0.4664.55

Node Version

No response

OS

Mac

Relevant log output

No response

[connorjclark]

I see the same (although I get ~4s with Fast 3g and 4x cpu). We'll have to take a look at why lantern gets this so wrong.

Thanks for the report!

[connorjclark]

In the meantime, you can trade load times for more accurate results by unchecking the Simulated Throttling option.

[pladaria]

Could this be caused by #14665 ?

https://codepen.io/pladaria/full/RwmKPGE
this page reports a LCP of ~10s when executed in mobile although the page loads in ~1s. What is causing this difference?

[adamraine]

@pladaria that is unrelated. Please see https://github.com/GoogleChrome/lighthouse#why-is-the-performance-score-so-low-it-looks-fine-to-me