Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
One thing I'm observing while trying to debug ProcedureV2 work is it's difficult to track an procedure as it's execution moves around the cluster. It would be nice if workers would include the PID (and maybe PPID) in the work they're performing. This probably means including this information as potential metadata in all of our RPC requests, and finding a way to log it uniformly.