I’m considering adding zipkin tracing in the RunTracker, which would render an entire pants run as a zipkin trace; it would re-use the WorkUnit tracking we do today, would at some point also add anything from the remote end of remote execution, and may go into more granularity into engine tasks… Any thoughts?
In particular, any thoughts on whether it should dump to the existing stats endpoint as another (optional) key, or whether it should post to a standalone single-purpose endpoint?
@happy-kitchen-89482 @enough-analyst-54434