Feature Request Type
Is your feature request related to a problem? Please describe.
Tracing issues in totality through the system (cumulus + whatever is talking to it) is impossible without a single unique-to-session-or-user identifier attached to each log statement. This becomes especially true when threads, forks, or other background tasks are involved.
Describe the solution you'd like
Support a user provided X-Trace-Id header or a generated TraceID for the session, that validates as UUIDish, during all logging of a user request.
OpenDCS's recent example: https://github.com/opendcs/opendcs/pull/1845/changes#diff-44cbaab8d6bc87468d995237fdafb4a6d9ca1fa8ba81202312b91bcb327a4375
Describe alternatives you've considered
Continue to never no which log statement is truly associated with a given larger session.
Additional context
A primary use-case for this TraceID is for, say HEC-RTS, to generate it at application, or watershed open, and use it for all calls to Cumulus, CDA, etc. If something doesn't go right, we can pull logs from each API, filtered by this TraceID, and now exactly what was happening for that session in everything.
Feature Request Type
Is your feature request related to a problem? Please describe.
Tracing issues in totality through the system (cumulus + whatever is talking to it) is impossible without a single unique-to-session-or-user identifier attached to each log statement. This becomes especially true when threads, forks, or other background tasks are involved.
Describe the solution you'd like
Support a user provided X-Trace-Id header or a generated TraceID for the session, that validates as UUIDish, during all logging of a user request.
OpenDCS's recent example: https://github.com/opendcs/opendcs/pull/1845/changes#diff-44cbaab8d6bc87468d995237fdafb4a6d9ca1fa8ba81202312b91bcb327a4375
Describe alternatives you've considered
Continue to never no which log statement is truly associated with a given larger session.
Additional context
A primary use-case for this TraceID is for, say HEC-RTS, to generate it at application, or watershed open, and use it for all calls to Cumulus, CDA, etc. If something doesn't go right, we can pull logs from each API, filtered by this TraceID, and now exactly what was happening for that session in everything.