bancolombia/distributed-performance-analyzer

Performance Analyzer代码仓库 https://github.com/bancolombia/distributed-performance-analyzer

Performance Analyzer是一款HTTP基准测试工具，能够从单个节点或分布式集群生成显著负载。它结合Elixir的能力，用于分析应用在不同并发场景下的行为。

快速参考**维护者：https://github.com/bancolombia/**%E9%97%AE%E9%A2%98%E5%8F%8D%E9%A6%88%E5%9C%B0%E5%9D%80%EF%BC%9A**https://github.com/bancolombia/distributed-performance-analyzer/issues 支持架构：**amd64, arm64

支持的标签及对应Dockerfile链接

• https://github.com/bancolombia/distributed-performance-analyzer/blob/main/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.6.3/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.6.2/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.6.1/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.6.0/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.5.0/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.4.1/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.4.0/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.3.1/Dockerfile
• https://github.com/bancolombia/distributed-performance-analyzer/blob/0.3.0/Dockerfile

安装

shell
mix do local.hex --force, local.rebar --force
mix do deps.clean --unused, deps.get, deps.compile
mix compile

基本使用

打开并编辑config/performance.exs文件进行配置。

elixir
import Config

config :perf_analyzer,
  url: "http://httpbin.org/get",
  request: %{
    method: "GET",
    headers: [{"Content-Type", "application/json"}],
    # body: ~s|'{"data": "value"}'| --> 如果不使用数据集值
    body: fn item ->
      # 用于数据集替换
      ~s|'{"data": #{Enum.random(1..10)}, "key": "#{item.columnName}"}'|
    end
  },
  execution: %{
    steps: 5,
    increment: 1,
    duration: 2000,
    constant_load: false,
    dataset: "/Users/sample.csv",
    # dataset: :none, --> 如果不使用数据集
    separator: ","
  },
  distributed: :none,
  jmeter_report: true

config :logger,
  level: :info

在上面的示例中，将执行一个包含5步、增量为50的测试：

1. 第1步：50并发
2. 第2步：100并发
3. 第3步：150并发
4. ...

每步将持续2秒。

运行

Docker方式

shell
docker run --rm -v <项目路径>/config:/app/config -v <项目路径>/dataset:/app/datasets bancolombia/distributed-performance-analyzer:latest

Shell方式

shell
iex -S mix
# 或
iex --sname node1@localhost -S mix

IEX执行

elixir
alias DistributedPerformanceAnalyzer.Domain.UseCase.ExecutionUseCase
ExecutionUseCase.launch_execution()

Elixir发布版

shell
mix release
_build/prod/rel/distributed_performance_analyzer/bin/distributed_performance_analyzer start

结果

每步执行后，将获得如下结果表格：

CSV格式

csv
concurrency, throughput, min latency (ms), mean latency (ms), max latency (ms), p90 latency (ms), p95 latency (ms), p99 latency (ms), http_mean_latency, http_max_latency, 2xx requests, 3xx requests, 4xx requests, 5xx requests, http_errors, total_errors, total_requests
1, 4.0, 27, 49.0, 71, 66.6, 68.8, 70.56, 49.0, 71, 2, 0, 0, 0, 0, 0, 2
2, 116.0, 12, 17.0, 33, 22.3, 24.0, 33.0, 17.0, 33, 58, 0, 0, 0, 0, 0, 58
3, 156.0, 12, 18.44, 28, 24.3, 25.15, 27.23, 18.44, 28, 78, 0, 0, 0, 0, 0, 78
4, 178.0, 12, 21.39, 50, 27.2, 47.0, 49.12, 21.39, 50, 89, 0, 0, 0, 0, 0, 89
5, 294.0, 12, 16.52, 32, 22.0, 24.4, 31.54, 16.52, 32, 147, 0, 0, 0, 0, 0, 147

JMeter格式

csv
timeStamp,elapsed,label,responseCode,responseMessage,threadName,dataType,success,failureMessage,bytes,sentBytes,grpThreads,allThreads,URL,Latency,IdleTime,Connect
1694523851175,71,sample,200,OK,#PID<0.443.0>,,true,,24,0,1,1,GET -> http://localhost:8080/wait/10,70,0.0,400
1694523851246,27,sample,200,OK,#PID<0.443.0>,,true,,24,0,1,1,GET -> http://localhost:8080/wait/10,27,0.0,2
1694523851317,15,sample,200,OK,#PID<0.443.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,15,0.0,0
1694523851317,15,sample,200,OK,#PID<0.444.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,15,0.0,400
1694523851332,22,sample,200,OK,#PID<0.443.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,22,0.0,0
1694523851332,22,sample,200,OK,#PID<0.444.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,22,0.0,0
1694523851354,17,sample,200,OK,#PID<0.443.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,17,0.0,0
1694523851354,17,sample,200,OK,#PID<0.444.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,17,0.0,0
1694523851371,20,sample,200,OK,#PID<0.443.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,20,0.0,0
1694523851371,20,sample,200,OK,#PID<0.444.0>,,true,,24,0,2,2,GET -> http://localhost:8080/wait/10,20,0.0,0

可根据需求比较关注的属性，例如并发与吞吐量、并发与平均延迟等指标。