Open anniesullie opened 4 years ago
One case that came to my mind and might point out the kind of overlap/combination with NEL is if we want to include in the new metric cases like:
Because I see that this is indeed a case that we could use NEL data, would we consider this case an occurance of abandonment?
My opinion in the case of errors is that the page is not really abandoned--the user received some kind of error message, which is kind of equivalent to an FCP on a well-behaved page. Would love to hear more opinions on this.
I just had a read of this and it's great, I think it'd be really valuable.
My personal opinion is that NEL should not include abandoned
as it's not a network error, it's a user action (assuming i am not missing something).
What I'd suggest as a possible addition to your proposal @anniesullie is that abandoned
is actually removed from NEL and replaced by this. Then also that there's a phase
or type
added to the abandonment report which would be e.g. during-request
or during-download
etc. - that would be a really useful discriminator to use in reports.
I also would strongly favour using the reporting API rather than Chrome User Experience reporting as it's (at least potentially) cross-browser and we can more easily merge the data with our existing NEL, CSP, Expect CT etc. reporting.
@neilstuartcraig a discussion about the "where" could this be exposed, can also be found here https://github.com/anniesullie/web-page-abandonment-api/issues/3#issuecomment-587951678
If we use a reporting API to report abandonment, there will be an overlap with the
abandoned
error in the Network Error Logging API. That API is similar to this one, but subtly different: