add conventions for log correlation

[zenmoto]

trafficstars

This OTEP is to add basic conventions for correlating logs with traces and resources by modifying existing logging formats.

[linux-foundation-easycla[bot]]

The committers are authorized under a signed CLA.

• :white_check_mark: David Poncelow (f8cc2cfea626c3b8cf009142729cba8f3f82561c, 0a724be725bbd3d3b916dea375272ffb2169939b)

[bogdandrutu]

@open-telemetry/specs-approvers @open-telemetry/specs-logs-approvers please review this.

[PatrikSteuer]

https://github.com/open-telemetry/opentelemetry-specification/pull/1283 changes the format of traceid, spanid, traceflags to trace_id, span_id, and trace_flags which seems to be related.

[kamalmarhubi]

it also seems that correlation context should be renamed to baggage? cf https://github.com/open-telemetry/opentelemetry-specification/issues/536 and https://github.com/open-telemetry/opentelemetry-specification/pull/857

[tigrannajaryan]

Most of the proposals in this PR are already part of the specification. This OTEP is actually someone out of sync / behind with the specification. I suggest that we close the OTEP and any additional clarifications necessary are submitted as issues/PRs against specification repo.

[mars64]

Most of the proposals in this PR are already part of the specification. This OTEP is actually someone out of sync / behind with the specification. I suggest that we close the OTEP and any additional clarifications necessary are submitted as issues/PRs against specification repo.

I know this is an aged pull here, but just wanted to point out that at least as of today the open-telemetry/opentelemetry-specification/specification/logs/overview.md doc still refers to this pull for definitions on log correlation behavior.

For posterity, I think what you're saying is that folks interested learning more/contributing to the log correlation spec should instead refer to and issue requests against the data-model.md doc/area, is that right? This area is also mentioned, but its confusing. It seems like removing the mention to this pull would be a quick win for anyone else interested in following along.

[tedsuo]

@tigrannajaryan the spec still refers to this OTEP: https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/logs/README.md#trace-context-in-legacy-formats

Should we go ahead and approve it?

[tigrannajaryan]

@tigrannajaryan the spec still refers to this OTEP: https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/logs/README.md#trace-context-in-legacy-formats

Should we go ahead and approve it?

This OTEP's ideas were incorporated in to the spec, so it served its purpose, however the spec slightly deviated from this OTEP in nuances. I will close this OTEP and it can stay for posterity. The spec probably needs slight re-wording to ensure it does not imply that the OTEP is the source of truth.

[tigrannajaryan]

@tigrannajaryan the spec still refers to this OTEP: https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/logs/README.md#trace-context-in-legacy-formats Should we go ahead and approve it?

This OTEP's ideas were incorporated in to the spec, so it served its purpose, however the spec slightly deviated from this OTEP in nuances. I will close this OTEP and it can stay for posterity. The spec probably needs slight re-wording to ensure it does not imply that the OTEP is the source of truth.

@zenmoto do you think you can condense this OTEP into 1-2 paragraphs that we can incorporate into the spec? Linking to the OTEP is a bit misleading (e.g. OTEP uses traceid, but we went with trace_id in the spec), so I think it is best to avoid linking to the OTEP.

[zenmoto]

The important parts of this OTEP have made their way into the spec. Closing this PR as it's outdated.