Best Practice Usage of Trace ID

[ethangodt]

• [X] This only affects the JavaScript OpenTelemetry library
• [ ] This may affect other libraries, but I would like to get opinions here first

Hello 👋

I'm working on a JS lib for my team and seeking advice about the best way to handle trace ids. I'm new to Open Telemetry, and my understanding at this point is that trace ids are used to correlate disparate events for enhanced visibility and discoverability.

For proof of concept work, I am currently testing with this vendor library which makes use of some default @open-telemetry instrumentations.

When I integrate this code into my page I'm correctly sending traces, but I'm noticing that there is a different trace id for each instrumentation. On one hand, it makes sense that fetch/ajax tracking have their own trace id (each request is traceable uniquely through a backend system), but in the case of other events (like all core web vitals, user click events, etc.) it seems desirable to correlate these with the same trace id.

Why do the libraries in this repo default to using a unique trace id in most cases? How should I think about this concept? Am I suppose to be supplying a trace id to these instrumentations at page load? Is there meaningful definition somewhere between a trace id and a session id?

Example: