Expose a batch rate limit request

[clarkds14]

There are situations where users would like to be able to send a batch request to minimize network roundtrips. The two types of requests that come to mind:

1. Sending rate limit requests for more than 1 domain/key/value tuples.
2. Sending rate limit requests for the same domain/key/value treated as if they were separate requests. There is no work around for this with the current API because summing requests could potentially put you over limit for the entire batch.

[mattklein123]

Can you propose an API change for discussion?

[clarkds14]

The gist would be something like:

rpc BatchShouldRateLimit(BatchRateLimitRequest) returns (BatchRateLimitResponse)

message BatchRateLimitRequest {
  repeated RateLimitRequest requests = 1;
}

message BatchRateLimitResponse {
  repeated RateLimitResponse responses = 1;
}

where the order of the responses matches the order of the requests. There are fields that can probably be pulled out of the inner message onto the top-level message and coalesced (e.g. response_headers_to_add) but otherwise this would achieve the goal.

[mattklein123]

Sure SGTM. Marking help wanted.