I know we probably only need databaseId and projectId at this moment, but in case we need more info that we cannot easily get from the request directly in the future, do we want to implement extract and match logic based on the proto? So that we don't need to change the code every time the routing annotation changes.
We implemented it for pure gRPC gapic client libraries, and this is how the autogenerated client for Bigtable look like(I know Bigtable does not really use it, but you get the idea).

I know we probably only need databaseId and projectId at this moment, but in case we need more info that we cannot easily get from the request directly in the future, do we want to implement extract and match logic based on the proto? So that we don't need to change the code every time the routing annotation changes. We implemented it for pure gRPC gapic client libraries, and this is how the autogenerated client for Bigtable look like(I know Bigtable does not really use it, but you get the idea).

@blakeli0 this is an interesting idea - I will file an issue to look into this as a refactor for the future