-
Notifications
You must be signed in to change notification settings - Fork 880
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove log instrumentation #720
Comments
I added the logger instrumentations 😄. The are definitely out of spec. I can maintain them outside of this repo if no one else wants them. |
Ah if there's a clear owner I'm less concerned :) It seems like an interesting feature but should it be migrated to use events instead of creating new spans? (I'm guessing it comes pre-events). Having events for warning messages on my failed span seems quite cool! FWIW, the reason I stumbled upon them is wanting to implement ID injection into logs. From what I recall, there was talk about speccing that out but that's been punted for now, but I think it's very important and want to implement it too so I may end up working in these modules too. |
Thanks for the link - yeah I sort of figured that's the reasoning. My thoughts are
I couldn't tell from glancing at the instrumentation if it adds any overhead when the config flag disables all log capture, hopefully it doesn't. Will leave it to you though on whether we switch these back to events or move them around or do anything at all :) |
My needs match the logs vision otep (capturing logs both inside and outside of spans). The logging instrumentation brings this functionality for me today, because I can map the spans back to our own logging data model in the exporter. But I agree this is totally out-of-spec, and is going to cause confusion in this repo, so agree with removing it, and we can bring it back once we have a logging data model. |
I randomly looked at the logging instrumentation and found
experimentalLogCaptureThreshold
. I'm not aware of anything close to this style of converting logs to ~0 duration spans in OpenTelemetry specs and suspect this is carried over from datadog. Does it make sense to remove this instrumentation to prevent confusion?The text was updated successfully, but these errors were encountered: