I want to trace RPC calls. However, looking through the code, this doesn't seem to be possible yet as there is no catch-all way to start and propagate traces.
Thus, I propose a Ktor-like Plugin System that allows you to add handlers throughout the service call lifecycle. As an initial implementation, adding pre- and post-RPC service call hooks would do just fine.
The user-facing API could look something like this:
Later, more hooks can be added, for example pre and post serialization, error handlers, ...
Hi, @SIMULATAN ! Thank you for logging the issue, we are considering the extension points (aka plugins) in our work, and they will be added, though I can't say when for now
I want to trace RPC calls. However, looking through the code, this doesn't seem to be possible yet as there is no catch-all way to start and propagate traces.
Thus, I propose a Ktor-like Plugin System that allows you to add handlers throughout the service call lifecycle. As an initial implementation, adding pre- and post-RPC service call hooks would do just fine.
The user-facing API could look something like this:
Later, more hooks can be added, for example pre and post serialization, error handlers, ...