✨ Add input, process and render durations to INP

[arichard-info]

Motivation

When analysing the INP, it's very useful to have details of the durations: the input, processing and rendering delay.

This information can be calculated from processingStart and processingEnd, which are returned by the PerformanceObserver.

It's a proposition, and the attributes can be renamed. And I don't know if there's equivalent work to be done to make this information available in the Datadog UI, but here's a first draft

Changes

Addition of the values

• interaction_to_next_paint_input
• interaction_to_next_paint_process
• interaction_to_next_paint_render
• interaction_to_next_paint_type.

Testing

• Local
• Staging
• Unit
• End to end

[bits-bot]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Antoine RICHARD seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}