rate-limit: make 429 response mapping configurable (#4879)

This commit enables the configuration of the mapping that translates 429
response code to a gRPC status code. By default, the Rate Limit filter
in Envoy translates a 429 HTTP response code to UNAVAILABLE as specified
in the gRPC mapping document. Google, however, recommends translating a
429 response to RESOURCE_EXHAUSTED. This commit provides a flag named
rate_limited_as_resource_exhausted in the RateLimit config which allows
users to explicitly specify whether they want 429 responses to be mapped
to RESOURCE_EXHAUSTED, while UNAVAILABLE remains the default.

References:
* https://github.com/grpc/grpc/blob/master/doc/http-grpc-status-mapping.md
* https://cloud.google.com/apis/design/errors#generating_errors

Signed-off-by: Venil Noronha <veniln@vmware.com>

Mirrored from https://github.com/envoyproxy/envoy @ f71a883b557a18cc418d4103b2f07a6780fc6576

envoy/config/filter/http/rate_limit/v2/rate_limit.proto

@@ -40,4 +40,9 @@ message RateLimit {
   // communication failure between rate limiting service and the proxy.
   // Defaults to false.
   bool failure_mode_deny = 5;
+
+  // Specifies whether a `RESOURCE_EXHAUSTED` gRPC code must be returned instead
+  // of the default `UNAVAILABLE` gRPC code for a rate limited gRPC call. The
+  // HTTP code will be 200 for a gRPC response.
+  bool rate_limited_as_resource_exhausted = 6;
 }