Add diagnostic and contextualizing data for metrics

[malchata]

Collecting field data for performance is a tall order to begin with, but analyzing it without contextual data can make interpreting and acting on field data much more difficult. In my opinion, I believe that adding the following diagnostic information for the following metrics could be helpful:

1. TTFB values should be supplemented with server timing API data (where available).
2. CLS values should be supplemented with data from the layout instability API.
3. LCP values should be supplemented with data from the largest contentful paint and element timing APIs.
4. FCP values should be supplemented with data from the paint timing API.
5. FID values should be supplemented with data from the user timing and long task APIs.

By exposing this diagnostic information in this library in a convenient fashion, developers may be more likely to record and act on contextual field data that may help them to better diagnose performance problems in the field.