Microsoft MVP성태의 닷넷 이야기
글쓴 사람
정성태 (techsharer at outlook.com)
홈페이지
첨부 파일

(시리즈 글이 2개 있습니다.)
.NET Framework: 2126. C# - 서버 측의 요청 제어 (Microsoft.AspNetCore.RateLimiting)
; https://www.sysnet.pe.kr/2/0/13362

.NET Framework: 2129. C# - Polly를 이용한 클라이언트 측의 요청 재시도
; https://www.sysnet.pe.kr/2/0/13376




C# - 서버 측의 요청 제어 (Microsoft.AspNetCore.RateLimiting)

닷넷의 경우 공식적으로 Microsoft.AspNetCore.RateLimiting을 통해 ASP.NET Core에서 사용할 수 있는 Middleware를 제공하고 있습니다.

Rate limiting middleware in ASP.NET Core
; https://learn.microsoft.com/en-us/aspnet/core/performance/rate-limit

위의 글을 정리해 사용법을 대충 알아볼까요? ^^

이를 위해 기본 ASP.NET Core Web API 프로젝트를 생성하고 Program.cs에 다음과 같이 원하는 요청 제어 코드를 추가할 수 있습니다.

using Microsoft.AspNetCore.RateLimiting;
using System.Threading.RateLimiting;

namespace WebApplication1
{
    public class Program
    {
        public static void Main(string[] args)
        {
            var builder = WebApplication.CreateBuilder(args);

            builder.Services.AddRateLimiter(_ => _
                .AddFixedWindowLimiter(policyName: "fixed", options =>
                {
                    options.PermitLimit = 4;
                    options.Window = TimeSpan.FromSeconds(5);
                    options.QueueProcessingOrder = QueueProcessingOrder.OldestFirst;
                    options.QueueLimit = 2;
                }));


            builder.Services.AddControllers();
            builder.Services.AddEndpointsApiExplorer();
            builder.Services.AddSwaggerGen();

            var app = builder.Build();

            app.UseRateLimiter();

            // Configure the HTTP request pipeline.
            if (app.Environment.IsDevelopment())
            {
                app.UseSwagger();
                app.UseSwaggerUI();
            }

            app.UseAuthorization();


            app.MapControllers();

            app.Run();
        }
    }
}

AddFixedWindowLimiter의 옵션 설정은 직관적으로 이해할 수 있습니다. 5초 당 4개의 요청을 허용하고, 요청이 밀리는 경우 최대 2개의 요청만 Queue에 보관하고 있습니다. 예를 들어, 동시 요청이 7개가 오면 4개는 먼저 처리되고, 2개는 Queue에 보관한 다음 1개는 거부를 하는 것입니다.

그리고 그런 설정을 대표하는 정책 이름을 "fixed"라고 지정하는데요, 해당 정책을 모든 Web API에 적용하고 싶다면 MapControllers와 연계해 RequireRateLimiting 확장 메서드를 다음과 같이 호출하면 됩니다.

// ...[생략]...

app.UseAuthorization();

app.MapControllers().RequireRateLimiting("fixed");

app.Run();

// ...[생략]...




자, 그럼 정말 Rate Limit가 잘 적용됐는지 테스트를 해볼까요? ^^ 이를 위해 다음과 같이 간략하게 콘솔 프로젝트를 만들고,

namespace ConsoleApp1;

internal class Program
{
    public const int MaxRequests = 7;
    static async Task<int> Main(string[] args)
    {
        var client = new HttpClient();

        var requests = new Task<HttpResponseMessage>[MaxRequests];
        for (int i = 0; i < MaxRequests; i++)
        {
            requests[i] = client.GetAsync("http://localhost:5203/WeatherForecast");
        }

        await Task.WhenAll(requests);

        for (int i = 0; i < MaxRequests; i++)
        {
            Console.WriteLine($"{i}: {requests[i].Result}");
        }

        return 0;
    }
}

실행하면 6개의 200 OK 응답과,

StatusCode: 200, ReasonPhrase: 'OK', Version: 1.1, Content: System.Net.Http.HttpConnectionResponseContent, Headers:
{
  Date: Thu, 01 Jun 2023 11:51:53 GMT
  Server: Kestrel
  Transfer-Encoding: chunked
  Content-Type: application/json; charset=utf-8
}

1개의 503 응답이 나옵니다.

StatusCode: 503, ReasonPhrase: 'Service Unavailable', Version: 1.1, Content: System.Net.Http.HttpConnectionResponseContent, Headers:
{
  Date: Thu, 01 Jun 2023 11:51:41 GMT
  Server: Kestrel
  Content-Length: 0
}

결국 4개의 요청은 바로 처리가 되었고, 2개는 Queue에 들어간 후 5초 지난 후에 처리가 되고, 나머지 1개는 요청 초과로 인해 503 오류 코드를 받은 것입니다.




이것으로 기본 사용법은 모두 설명했습니다. 이하 약간의 세부 사항을 들어가 보면!

특정 Web API만을 제외하고 싶다면 DisableRateLimiting 특성을 사용할 수 있습니다. 위의 예제 코드에서는 "app.MapControllers().RequireRateLimiting("fixed");" 코드로 전체 서비스에 적용했는데, 특정 API만 제외하고 싶다면 다음과 같이 처리할 수 있습니다.

using Microsoft.AspNetCore.Mvc;
using Microsoft.AspNetCore.RateLimiting;

namespace WebApplication1.Controllers;

[ApiController]
[Route("[controller]")]
public class WeatherForecastController : ControllerBase
{
    // ...[생략]...

    [DisableRateLimiting]
    [HttpGet(Name = "GetWeatherForecast")]
    public IEnumerable<WeatherForecast> Get()
    {
        // ...[생략]...
    }
}

혹은, 반대로 RequireRateLimiting("fixed");" 호출을 제거해 모든 API에 대한 요청 제어를 하지 않고, 특정 API만 적용하고 싶다면 EnableRateLimiting 특성을 사용하면 됩니다.

[EnableRateLimiting("fixed")] // class 수준에도 적용 가능
[HttpGet(Name = "GetWeatherForecast")]
public IEnumerable<WeatherForecast> Get()
{
    // ...[생략]...
}

그 외에, 이제까지 설명한 Limiter는 AddFixedWindowLimiter 메서드의 이름에서 알 수 있듯이 "Fixed Window" 유형이었는데요, 부가적으로 다음과 같은 유형의 요청 제어도 제공하고 있습니다.


위의 기능들을 이용하면, 여러분도 OpenAI와 같은 식의 요청 제어를 할 수 있을 것입니다. ^^

(첨부 파일은 이 글의 예제 코드를 포함합니다.)




[이 글에 대해서 여러분들과 의견을 공유하고 싶습니다. 틀리거나 미흡한 부분 또는 의문 사항이 있으시면 언제든 댓글 남겨주십시오.]







[최초 등록일: ]
[최종 수정일: 6/1/2023]

Creative Commons License
이 저작물은 크리에이티브 커먼즈 코리아 저작자표시-비영리-변경금지 2.0 대한민국 라이센스에 따라 이용하실 수 있습니다.
by SeongTae Jeong, mailto:techsharer at outlook.com

비밀번호

댓글 작성자
 




[1]  2  3  4  5  6  7  8  9  10  11  12  13  14  15  ...
NoWriterDateCnt.TitleFile(s)
13960정성태7/3/2025440개발 환경 구성: 752. ProcDump - C/C++ 예외 코드 필터를 지정한 덤프 생성
13959정성태6/25/20251289오류 유형: 966. Ubuntu - ping: connect: Network is unreachable
13958정성태6/21/20251615닷넷: 2339. C# - Phi-4-multimodal 모델의 GPU 가속 방법 (ORT 사용)파일 다운로드1
13957정성태6/20/20251470닷넷: 2338. C# / Foundry Local - Phi-4-multimodal 모델을 사용하는 방법
13956정성태6/19/20251653개발 환경 구성: 751. Triton Inference Server의 Python Backend 프로세스
13955정성태6/18/20251777오류 유형: 965. Hugging Face 모델 다운로드 시 "requests.exceptions.HTTPError: 401 Client Error: Unauthorized for url: ..." 오류
13954정성태6/18/20251601닷넷: 2337. C# - Hugging Face에 공개된 LLM 모델을 Foundry Local에서 사용하는 방법파일 다운로드1
13953정성태6/16/20251425스크립트: 78. 파이썬 - 소스 코드의 파일 경로를 지정한 모듈 로드
13952정성태6/15/20251766닷넷: 2336. C# - IValueTaskSource로 인해 주의가 필요한 ValueTask 호출파일 다운로드1
13951정성태6/15/20251629오류 유형: 964. Outlook - 일정이 "You cannot make changes to contents of this read-only folder." 오류 메시지로 삭제가 안 되는 경우
13950정성태6/12/20252126닷넷: 2335. C# - 간단하게 구현해 보는 IValueTaskSource 예제파일 다운로드1
13949정성태6/11/20251831오류 유형: 963. SignTool - "Error: SignerSign() failed." (-2146869243/0x80096005)
13948정성태6/10/20251429오류 유형: 962. 파이썬 - Linux 환경 + TCP 서버 소켓을 사용하는 프로세스 종료 후 재실행하는 경우 "OSError: [Errno 98] Address already in use" 오류 발생
13947정성태6/9/20251910개발 환경 구성: 750. 파이썬 - Azure App Service에 응용 프로그램 배포 후의 환경
13946정성태6/9/20252019개발 환경 구성: 749. 파이썬 - Azure App Service에 응용 프로그램 배포하기 전의 환경
13945정성태6/7/20251728오류 유형: 961. 파이썬 + conda - mysqlclient 사용 시 "NameError: name '_mysql' is not defined" 에러
13944정성태6/7/20255473오류 유형: 960. The trust relationship between this workstation and the primary domain failed. - 네 번째 이야기
13943정성태6/6/20252139개발 환경 구성: 748. Windows + Foundry Local - 로컬에서 AI 모델 활용 [1]
13942정성태6/5/20251692오류 유형: 959. winget 설치 시 "0x80d02002 : unknown error"
13941정성태6/2/20251528닷넷: 2334. C# - cpuid 명령어를 이용한 CPU 제조사 문자열 가져오기파일 다운로드1
13940정성태6/1/20251969C/C++: 188. C++의 32비트 + Release 어셈블리 코드를 .NET으로 포팅할 때 주의할 점파일 다운로드1
13939정성태5/29/20252478오류 유형: 958. NVIDIA Triton Inference Server - version `GLIBCXX_3.4.32' not found (required by /opt/tritonserver/backends/python/triton_python_backend_stub)
13938정성태5/29/20251931개발 환경 구성: 747. 파이썬 - WSL/docker에 구성한 Triton 예제 개발 환경
13937정성태5/24/20251877개발 환경 구성: 746. Windows + WSL2 환경에서 (tensorflow 등의) NVIDIA GPU 인식
13936정성태5/23/20251782개발 환경 구성: 745. Linux / WSL 환경에 Miniconda 설치하기
[1]  2  3  4  5  6  7  8  9  10  11  12  13  14  15  ...