Open tunetheweb opened 1 year ago
The way I see it (and reflecting some internal team conversations with @mmocny and @tdresser), bad responsiveness can be caused by:
So for someone that's debugging INP, it would be best to start by understanding which kind of problem they have out of the (3), and see if correlating the exact event-timing entry with LoAF is what would help them with that.
The main reason we're adding LoAF is to better explain INP entries. Adding an explicit link between an INP entry and its LoAF sounds logical.
@noamr based on what you said it seems that more important than linking LoAF to INP would be to breakdown the 3 phases of INP (Input delay, Processing time and Presentation delay). Is it possible to do that with the current APIs? I know we can see that in chrome devtools but It would be good to track it in RUM
@noamr based on what you said it seems that more important than linking LoAF to INP would be to breakdown the 3 phases of INP (Input delay, Processing time and Presentation delay). Is it possible to do that with the current APIs? I know we can see that in chrome devtools but It would be good to track it in RUM
Event timing gives you precisely that breakdown, yes.
Been discussed a bit internally with the Chrome team, but here's a public issue to track and discuss this.
Currently LoAF is a separate view of Performance Timeline entries, meaning there is no direct link to an INP entry. As one of the primary reasons for this new API was as a tool for improving responsiveness (as measured by INP) should it be easier to get the LoAF entries for an INP entry?
Currently, this requires quite a lot of processing (example WhyNP script).