data-plane-api/envoy/service/ratelimit/v2/rls.proto

syntax = "proto3";

package envoy.service.ratelimit.v2;

option java_outer_classname = "RlsProto";
option java_multiple_files = true;
option java_package = "io.envoyproxy.envoy.service.ratelimit.v2";
option java_generic_services = true;

import "envoy/api/v2/core/base.proto";
import "envoy/api/v2/ratelimit/ratelimit.proto";

import "validate/validate.proto";

// [#protodoc-title: Rate Limit Service (RLS)]

service RateLimitService {
  // Determine whether rate limiting should take place.
  rpc ShouldRateLimit(RateLimitRequest) returns (RateLimitResponse) {
  }
}

// Main message for a rate limit request. The rate limit service is designed to be fully generic
// in the sense that it can operate on arbitrary hierarchical key/value pairs. The loaded
// configuration will parse the request and find the most specific limit to apply. In addition,
// a RateLimitRequest can contain multiple "descriptors" to limit on. When multiple descriptors
// are provided, the server will limit on *ALL* of them and return an OVER_LIMIT response if any
// of them are over limit. This enables more complex application level rate limiting scenarios
// if desired.
message RateLimitRequest {
  // All rate limit requests must specify a domain. This enables the configuration to be per
  // application without fear of overlap. E.g., "envoy".
  string domain = 1;

  // All rate limit requests must specify at least one RateLimitDescriptor. Each descriptor is
  // processed by the service (see below). If any of the descriptors are over limit, the entire
  // request is considered to be over limit.
  repeated api.v2.ratelimit.RateLimitDescriptor descriptors = 2;

  // Rate limit requests can optionally specify the number of hits a request adds to the matched
  // limit. If the value is not set in the message, a request increases the matched limit by 1.
  uint32 hits_addend = 3;
}

// A response from a ShouldRateLimit call.
message RateLimitResponse {
  enum Code {
    // The response code is not known.
    UNKNOWN = 0;

    // The response code to notify that the number of requests are under limit.
    OK = 1;

    // The response code to notify that the number of requests are over limit.
    OVER_LIMIT = 2;
  }

  // Defines an actual rate limit in terms of requests per unit of time and the unit itself.
  message RateLimit {
    enum Unit {
      // The time unit is not known.
      UNKNOWN = 0;

      // The time unit representing a second.
      SECOND = 1;

      // The time unit representing a minute.
      MINUTE = 2;

      // The time unit representing an hour.
      HOUR = 3;

      // The time unit representing a day.
      DAY = 4;
    }

    // The number of requests per unit of time.
    uint32 requests_per_unit = 1;

    // The unit of time.
    Unit unit = 2;
  }

  message DescriptorStatus {
    // The response code for an individual descriptor.
    Code code = 1;

    // The current limit as configured by the server. Useful for debugging, etc.
    RateLimit current_limit = 2;

    // The limit remaining in the current time unit.
    uint32 limit_remaining = 3;
  }

  // The overall response code which takes into account all of the descriptors that were passed
  // in the RateLimitRequest message.
  Code overall_code = 1;

  // A list of DescriptorStatus messages which matches the length of the descriptor list passed
  // in the RateLimitRequest. This can be used by the caller to determine which individual
  // descriptors failed and/or what the currently configured limits are for all of them.
  repeated DescriptorStatus statuses = 2;

  // [#next-major-version: rename to response_headers_to_add]
  repeated api.v2.core.HeaderValue headers = 3;

  // A list of headers to add to the request when forwarded
  repeated api.v2.core.HeaderValue request_headers_to_add = 4;
}