zipkin4net icon indicating copy to clipboard operation
zipkin4net copied to clipboard

Add propagation using W3C Trace Context headers

Open bgrainger opened this issue 6 years ago • 3 comments

W3C Distributed Trace Context Community Group has an editor's draft of a propagation format for distributed trace context: https://w3c.github.io/distributed-tracing/report-trace-context.html

Would you be interested in receiving a PR that added an IPropagation<K> implementation that (de)serializes traces in the W3C format?

bgrainger avatar Apr 12 '18 17:04 bgrainger

Usually we let PRs come in response to user demand.. are you a user?

Regardless, something like this if done in a way safe to the codebase could be helpful in anticipation of demand, for example to allow interop testing. I'd defer to @fedj, but if the code was pluggable and allows for revisions to the spec to not break production users, seems helpful and something we could consider bending "rule of three" on.

codefromthecrypt avatar Apr 13 '18 01:04 codefromthecrypt

are you a user?

Trying to become one; see https://github.com/openzipkin/zipkin4net/pull/198#issuecomment-380841943

if the code was pluggable and allows for revisions to the spec to not break production users

This is probably the biggest reason not to add it at this point (and something I should have mentioned in the original issue); it may be sensible to wait for the spec to be finalised.

I'm happy to keep the code in my own library for now (https://github.com/Faithlife/FaithlifeTracing/pull/1/files#diff-49c766b2446f8555e29e349a8525f2d3) but also contribute it here (to help others) once the W3C recommendation starts getting traction.

bgrainger avatar Apr 13 '18 03:04 bgrainger

Trying to become one; see #198 (comment) https://github.com/openzipkin/zipkin4net/pull/198#issuecomment-380841943

thanks!

if the code was pluggable and allows for revisions to the spec to not break production users

This is probably the biggest reason not to add it at this point (and something I should have mentioned in the original issue); it may be sensible to wait for the spec to be finalised.

I'm happy to keep the code in my own library for now ( https://github.com/Faithlife/FaithlifeTracing/pull/1/files#diff- 49c766b2446f8555e29e349a8525f2d3) but also contribute it here (to help others) once the W3C recommendation starts getting traction.

FYI I plan to make something in brave (java tracer), as we have a propagation plugin anyway. So, whether here or somewhere else it would be good to have means to test this. Otherwise the format may have flaws we don't have a chance to help fix

codefromthecrypt avatar Apr 13 '18 03:04 codefromthecrypt