Initial work to switch to tracing#2185
Conversation
|
Does your work include emitting traces that follow the OpenTelemetry db convention? https://opentelemetry.io/docs/reference/specification/trace/semantic_conventions/database/ That would just be awesome. |
|
@CosmicHorrorDev do you mind rebasing this please? |
|
@gitmalong I doubt it. I kept the messages as is since the focus of this PR is just to switch the logging framework from @abonander Finished rebasing. Should be good for a review |
|
This looks good! I remember looking into |
* Add tracing dep * Switch over basic events * Switch over dynamically enabled events * Fix missing SocketAddr formatting * More format fixing * refactor: Apply tracing changes to new crate structure
* Add tracing dep * Switch over basic events * Switch over dynamically enabled events * Fix missing SocketAddr formatting * More format fixing * refactor: Apply tracing changes to new crate structure
|
|
|
I would leave a comment on #2177 instead since that's the pseudo tracking issue for the tracing switch |
Resolves #2177
cc @taylor1791 (Sorry for the slight delay. Forgot it was Halloween weekend)
This is the initial work to switch from
logtotracingRoadbumps
event!()andenabled!()macros which led to some ugly macroslogworld (e.g.RUST_LOG=debugwithenv_logger), so bothtracingandlogare checked for dynamic bitslogfeature oftracingpulls inloganyways, so it doesn't add any extra depsFuture Work
It would probably be worthwhile to have some spans setup for tracking calls throughout the library, but I'll leave that to someone else since life is getting busy again