@thomasrossignolMay 06.2021 — #Humm... actually it depends where the Core Web Vitals are measured from.
- if your measurement is done from a vantage point which is located in the cloud / a datacenter (test platforms like PageSpeed Insights or webpagetest.org), it is unlikely that it is going to be affected by an increased internet latency.
- if your measurement is done from the actual users (assuming your are using a real user monitoring solution or GoogleSearchConsole), in this case the measurement will be impacted by the users slow internet connection or increased latency to your platform or webserver(s).
If network latency is high between the measurement point (actual user or vantage point for a service like PageSpeed Insights), then two of the 3 Core Web Vitals may be affected.
The main impact will be on LCP (Largest Contentful Paint) which reflects the loading speed and hence that metric will be impacted by the time needed to establish the TCP and TLS sessions and the transfer time for each of the resources of the rendering path.
To understand this in more details, I can point you to two articles: