Paulo Valente
Paulo Valente
I like the idea of having a default implementation + overridable behaviour
As a point of reference, I'll work to publish events so we can export all metrics that grpc_prometheus currently exports
Moving this to 0.7 because it's been too long since 0.5 and there are people depending on us releasing an update to Hex.
Closing this as stale for now
We could point to it in the README or on the Telemetry module in that case!
Maybe we should add a specific telemetry event for this?
> @polvalente why mix client request/response logic and telemetry? Those are separate concerns. Just a suggestion. I didn't perceive this as being just a client/server logic problem from the descriptions...
Perhaps we could introduce something close to the Client.Stream struct that plays into client streaming. Then we could pass either one of those to GRPC.Stub.cancel Furthermore, the struct itself could...
@hackvan I think that's perfect. We can put that in `GRPC.Status` and add something like "See `GRPC.Status` for more details on possible statuses" in `GRPC.Error`
I think we could provide this macro in GRPC.RPCError together with the better documentation: ```elixir defmacro is_rpc_error(e, status) when is_struct(e, GRPC.RPCError) and e.status == status ``` Which would make it...