Question: why we use spanId and rootSpanId to correlate profiles with traces

[steverao]

Hi everyone!

I have a question about the profile module in the codebase. Specifically, I would like to understand why we use spanId and rootSpanId to correlate profiles with traces in Datadog's dd-trace-java repository.

In my opinion, there are several approaches to correlating profiles with traces:

1. Use traceId and spanId to correlate each span with the related profile data.
2. Use traceId to correlate the entire trace with the related profile data.

However, I noticed that we use spanId and rootSpanId in the dd-trace-java project. Is there any deeper consideration behind this choice?

I hope to receive your response. Thank you very much!