pdpv0: tracing with requestIds

[BigLep]

Done Criteria

Curio accepts or generates a trace/request id for each request (especially "write" requests) and then logs that id for all work associated with the request (including asynchronous tasks).

Why Important

The general debugging superpower that get from traceability, both for root causing problems and analyzing performance.

User/Customer

Primarily for maintainers to help with their debugging and performance analysis.

Notes

1. This shares some similarity to pdpv0: self-service way for maintainers to get SP logs #712, but it was broken out because it's presumably a much bigger changes since it involves changes in the Curio codebase (and likely DB schema changes as well).
2. There is also a shorter-term quick win step of just logging when we receive a HTTP request with "request-id" or "trace-id" headers: pdpv0: log if receive a request with "request-id" or "trace-id" headers #729

[BigLep]

I'll close this issue since it hasn't turned out to be needed in the last 1.5 months and I don't expect it will be needed for GA. We'll reopen if the practical need emerges.