search

vercel / next.js

The React Framework
https://nextjs.org
MIT License
126.98k stars 26.98k forks source link

Test the built -in server is very slow, use ApacheBench. #50852

Open zhang-wenchao opened 1 year ago

zhang-wenchao commented 1 year ago

Verify canary release

Provide environment information

$ npx --no-install next info

    Operating System:
      Platform: linux
      Arch: x64
      Version: #1 SMP PVE 5.15.107-2 (2023-05-10T09:10Z)
    Binaries:
      Node: 18.16.0
      npm: 9.5.1
      Yarn: 1.22.19
      pnpm: N/A
    Relevant packages:
      next: 13.4.5-canary.6
      eslint-config-next: 13.4.4
      react: 18.2.0
      react-dom: 18.2.0
      typescript: 5.0.4

Which area(s) of Next.js are affected? (leave empty if unsure)

Middleware / Edge (API routes, runtime)

Link to the code that reproduces this issue or a replay of the bug

npx create-next-app@latest

To Reproduce

npx create-next-app@latest

use default config.

npm run build
npm run start

Describe the Bug

Very slow response speed and concurrency.

$ ab -n 10000 -c 100 http://192.168.100.10:3000/   
This is ApacheBench, Version 2.3 <$Revision: 1903618 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 192.168.100.10 (be patient)
Completed 1000 requests
Completed 2000 requests
Completed 3000 requests
Completed 4000 requests
Completed 5000 requests
Completed 6000 requests
Completed 7000 requests
Completed 8000 requests
Completed 9000 requests
Completed 10000 requests
Finished 10000 requests

Server Software:        
Server Hostname:        192.168.100.10
Server Port:            3000

Document Path:          /
Document Length:        16160 bytes

Concurrency Level:      100
Time taken for tests:   11.076 seconds
Complete requests:      10000
Failed requests:        0
Total transferred:      164740000 bytes
HTML transferred:       161600000 bytes
Requests per second:    902.82 [#/sec] (mean)
Time per request:       110.764 [ms] (mean)
Time per request:       1.108 [ms] (mean, across all concurrent requests)
Transfer rate:          14524.49 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.2      0       3
Processing:    37  110  14.5    107     257
Waiting:       32  110  14.4    107     257
Total:         37  110  14.5    107     257

Percentage of the requests served within a certain time (ms)
  50%    107
  66%    111
  75%    115
  80%    118
  90%    126
  95%    135
  98%    145
  99%    160
 100%    257 (longest request)

Expected Behavior

Real：

Requests per second:    856.75 [#/sec] (mean)

Estimate

Requests per second:    5000.75 [#/sec] (mean)

Which browser are you using? (if relevant)

No response

How are you deploying your application? (if relevant)

next start

zhang-wenchao commented 1 year ago

Compare with a Golang Gin service

Requests per second: 14790.47 [#/sec] (mean)

$ ab -n 10000 -c 1000 http://0.0.0.0:8080/v1/util/time                              
This is ApacheBench, Version 2.3 <$Revision: 1903618 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 0.0.0.0 (be patient)
Completed 1000 requests
Completed 2000 requests
Completed 3000 requests
Completed 4000 requests
Completed 5000 requests
Completed 6000 requests
Completed 7000 requests
Completed 8000 requests
Completed 9000 requests
Completed 10000 requests
Finished 10000 requests

Server Software:        
Server Hostname:        0.0.0.0
Server Port:            8080

Document Path:          /v1/util/time
Document Length:        53 bytes

Concurrency Level:      1000
Time taken for tests:   0.676 seconds
Complete requests:      10000
Failed requests:        0
Total transferred:      4660000 bytes
HTML transferred:       530000 bytes
Requests per second:    14790.47 [#/sec] (mean)
Time per request:       67.611 [ms] (mean)
Time per request:       0.068 [ms] (mean, across all concurrent requests)
Transfer rate:          6730.82 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0   29   6.5     30      46
Processing:    14   35   9.4     34      65
Waiting:        1   25   8.2     24      58
Total:         33   64   5.9     64      82

Percentage of the requests served within a certain time (ms)
  50%     64
  66%     65
  75%     67
  80%     67
  90%     72
  95%     76
  98%     77
  99%     78
 100%     82 (longest request)
zhang-wenchao commented 1 year ago

use sample api route

Requests per second: 908.80 [#/sec] (mean)

import { NextResponse } from 'next/server';

export async function GET() {
    return NextResponse.json({ "test":"test" });
}
$ ab -n 10000 -c 1000 http://192.168.100.10:3000/api
This is ApacheBench, Version 2.3 <$Revision: 1903618 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 192.168.100.10 (be patient)
Completed 1000 requests
Completed 2000 requests
Completed 3000 requests
Completed 4000 requests
Completed 5000 requests
Completed 6000 requests
Completed 7000 requests
Completed 8000 requests
Completed 9000 requests
Completed 10000 requests
Finished 10000 requests

Server Software:        
Server Hostname:        192.168.100.10
Server Port:            3000

Document Path:          /api
Document Length:        15 bytes

Concurrency Level:      1000
Time taken for tests:   11.003 seconds
Complete requests:      10000
Failed requests:        0
Total transferred:      2170000 bytes
HTML transferred:       150000 bytes
Requests per second:    908.80 [#/sec] (mean)
Time per request:       1100.347 [ms] (mean)
Time per request:       1.100 [ms] (mean, across all concurrent requests)
Transfer rate:          192.59 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    2   6.9      0      32
Processing:    43 1056 189.7   1003    2045
Waiting:       10 1055 189.7   1003    2045
Total:         43 1058 193.1   1004    2059

Percentage of the requests served within a certain time (ms)
  50%   1004
  66%   1026
  75%   1059
  80%   1115
  90%   1224
  95%   1458
  98%   1828
  99%   1947
 100%   2059 (longest request)
zhang-wenchao commented 1 year ago

I think this is a problem, because this performance is extremely low.

zhang-wenchao commented 1 year ago

The simple API request should have a faster speed, and the results are not. There must be any problems here.

zhang-wenchao commented 1 year ago

I think I know where the problem is and the solution to the problem.

I started two services with the same configuration, just different port numbers different.

config nginx.conf

 # 添加socket转发的代理
    upstream socket_proxy {
        # 转发的目的地址和端口
        server 192.168.100.10:3000;
        server 192.168.100.10:3001;
    }

    # 提供转发的服务，即访问localhost:9001，会跳转至代理socket_proxy指定的转发地址
    server {
       listen 2200;
       proxy_pass socket_proxy;
    }
$ ab -n 10000 -c 200 http://192.168.10.1:2200/api
This is ApacheBench, Version 2.3 <$Revision: 1903618 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 192.168.10.1 (be patient)
Completed 1000 requests
Completed 2000 requests
Completed 3000 requests
Completed 4000 requests
Completed 5000 requests
Completed 6000 requests
Completed 7000 requests
Completed 8000 requests
Completed 9000 requests
Completed 10000 requests
Finished 10000 requests

Server Software:
Server Hostname:        192.168.10.1
Server Port:            2200

Document Path:          /api
Document Length:        15 bytes

Concurrency Level:      200
Time taken for tests:   5.724 seconds
Complete requests:      10000
Failed requests:        0
Total transferred:      2170000 bytes
HTML transferred:       150000 bytes
Requests per second:    1746.90 [#/sec] (mean)
Time per request:       114.488 [ms] (mean)
Time per request:       0.572 [ms] (mean, across all concurrent requests)
Transfer rate:          370.19 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.7      0       7
Processing:     3  113 101.0    121     320
Waiting:        2  112 101.0    120     320
Total:          3  113 101.0    124     320

Percentage of the requests served within a certain time (ms)
  50%    124
  66%    192
  75%    200
  80%    211
  90%    244
  95%    255
  98%    285
  99%    305
 100%    320 (longest request)

Nearly double performance improvement.

naseef0 commented 7 months ago

@zhang-wenchao Im also facing similar issue when i test load with 30 users, app getting 502.