> gRPC (currently) always uses 200 for RPC errors I wanted to confirm that servers should return a 200 HTTP status code regardless of gRPC status. I believe this is the case but I didn't find https://github.com/grpc/grpc/blob/master/doc/PROTOCOL-HTTP2.md to be explicit about it.
On Monday, October 7, 2024 at 1:19:03 PM UTC-4 Eric Anderson wrote: > google/rpc/code.proto > <https://github.com/googleapis/googleapis/blob/master/google/rpc/code.proto> > has the HTTP mappings. Those are REST-centric and are for converting *to* > HTTP status code. Multiple of them are obviously wrong for gRPC (e.g., > NOT_FOUND→404) and multiple of them are... weak... mappings even with REST. > They aren't used for gRPC as gRPC (currently) always uses 200 for RPC > errors. (I would be interested in changing that, in particular for > UNIMPLEMENTED, but it hasn't been a problem.) > > http-grpc-status-mapping.md is the opposite. It is HTTP → gRPC status > code. HTTP 429 could be mapped to RESOURCE_EXHAUSTED, but I don't think > UNAVAILABLE is *bad*. There was some discussion on the PR about this > specifically after it was merged: > https://github.com/grpc/grpc/pull/4955#pullrequestreview-6858335 . The > current mapping was trying to be "good enough," and was thought to be the > first pass. I'm quite surprised it has been 8 years without *any* change. > > On Mon, Oct 7, 2024 at 7:09 AM Trustin Lee <tru...@gmail.com> wrote: > >> Hi, >> >> According to http-grpc-status-mapping.md >> <https://github.com/grpc/grpc/blob/v1.66.2/doc/http-grpc-status-mapping.md>, >> HTTP 429 Too Many Requests is translated into gRPC UNAVAILABLE rather than >> RESOURCE_EXHAUSTED. However, at some point in the past, statuscodes.md >> <https://github.com/grpc/grpc/blob/v1.23.1/doc/statuscodes.md> said >> RESOURCE_EXHAUSTED. Is there any reason the gRPC team changed the >> recommended mapping for HTTP 429 status? Why is UNAVAILABLE is better than >> RESOURCE_EXHAUSTED for HTTP 429? >> >> I'm asking this question not only to solve my own curiosity, but also to >> give a better answer to the bug report >> <https://github.com/line/armeria/issues/5928> for Armeria's gRPC status >> code mapping that merely follows what grpc-java does. >> >> Thanks, >> T >> >> -- >> You received this message because you are subscribed to the Google Groups >> "grpc.io" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to grpc-io+u...@googlegroups.com. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/grpc-io/24702a70-915d-41d5-a9a1-e477bfaf8a05n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/grpc-io/24702a70-915d-41d5-a9a1-e477bfaf8a05n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "grpc.io" group. To unsubscribe from this group and stop receiving emails from it, send an email to grpc-io+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/grpc-io/805eb44e-fa31-481f-8231-71071e87e99fn%40googlegroups.com.
