Support for GLS and comment-based tracing #76
Unanswered
ian-axelrod-sl
asked this question in
Q&A
Replies: 1 comment 1 reply
|
I believe the GLS feature should be included in the first version. Although it hasn't been developed yet, it has been on our high-priority to-do list for a long time. As for comment-based tracing, I don't think we will proactively include it in the first version, but if a solid community contribution comes along, it should be perfectly acceptable. Thanks! |
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Hi,
I would like to know if you are planning to implement a GLS-like solution, similar to orchestrion, to support tracing in codebases that have not followed the practice of including a
Contextin many function arguments.Context (heh...): I want to add tracing to a collection of large codebases that have a fairly deep call stack for functions, many of which are not currently using a
Context. It will take ages to refactor this codebase to add the contexts, as I'll need to add them all the way up the stack and then fix broken tests. I'll need to do this for each function (there are tens-to-hundreds per codebase) and each codebase (there are upwards of 10 codebases). I can use orchestrion temporarily to handle this, but I want a vendor-agnostic solution. Any plans here?Comment-based tracing: Orchestrion has the idea of marking a function for tracing using
//dd:spancomments. This has been invaluable for quickly adding support for tracing to functions without significantly altering structs / interfaces / parameters. The tracing library I work with at the moment is an abstraction over opentelemetry and datadog, and does not use a global mutex. I am required to pass the tracer abstraction to the function somehow (similar to context, but in this case I can at least pass in a struct / in the receiver) so that I can create new spans or add span attributes/baggage.//dd:spanhas allowed me to get around this issue for functions where we do not need to add extra span tags at runtime; we just want a new span. This is most of the functions. It would be great to have a similar mechanism in otel. This has the additional benefit that you can manage span lifecycle for the developer by defering for them. Thoughts on this?Thanks, fellas! And great project!
All reactions