[speedgoose]

> JSON objects are large and field names are repetitive

I used to write protocol buffer stuff for this reason. But I realized after some time that compressed json is almost as good if not better depending on the data, and a lot simpler and nicer to use. You can consider to pre-share a dictionary if you want to compress always the same tiny messages. Of course json + compression is a bit more cpu intensive than protocol buffers but it's not having an impact on anything in most use cases.

[ericbarrett]

I think your instinct to reach for the straightforward solution is good. gRPC has advantages, but it also comes with complexity since you have to bring all the tooling along. And the CPU burden of (de-)serializing JSON is a very different story than when Protobufs were developed in 2001.