Back to Details

testing-performance

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Performance Testing

性能测试

Test application speed, responsiveness, stability, and scalability under various load conditions.
测试应用在不同负载条件下的速度、响应性、稳定性以及可扩展性。

When to use me

适用场景

Use this skill when:
  • Preparing for traffic spikes or seasonal loads
  • Testing application scalability
  • Identifying performance bottlenecks
  • Validating SLAs (Service Level Agreements)
  • Comparing performance before/after changes
  • Capacity planning and infrastructure sizing
  • Ensuring user experience under load
当你有以下需求时可使用本技能：
  • 为流量高峰或季节性负载做准备
  • 测试应用可扩展性
  • 识别性能瓶颈
  • 验证SLA（服务水平协议）
  • 对比变更前后的性能表现
  • 容量规划与基础设施规格选型
  • 保障负载下的用户体验

What I do

功能说明

  • Load testing: Simulate expected user traffic
  • Stress testing: Find breaking points and limits
  • Spike testing: Test sudden traffic surges
  • Endurance testing: Check for memory leaks over time
  • Scalability testing: Verify horizontal/vertical scaling
  • Configuration testing: Optimize performance settings
  • Benchmarking: Compare against baseline metrics
  • 负载测试：模拟预期用户流量
  • 压力测试：找到系统断点与性能上限
  • 峰值测试：测试突发流量激增场景
  • 耐久性测试：长时间运行检测内存泄漏
  • 可扩展性测试：验证水平/垂直扩容效果
  • 配置测试：优化性能配置项
  • 基准测试：与基线指标做对比

Examples

示例

bash
undefined
bash
undefined

Load testing tools

Load testing tools

npx autocannon -c 100 -d 60 https://app.example.com wrk -t12 -c400 -d30s https://app.example.com k6 run script.js # Grafana k6 jmeter -n -t testplan.jmx -l results.jtl
npx autocannon -c 100 -d 60 https://app.example.com wrk -t12 -c400 -d30s https://app.example.com k6 run script.js # Grafana k6 jmeter -n -t testplan.jmx -l results.jtl

Performance monitoring

Performance monitoring

npm run test:perf # Custom performance suite lighthouse https://app.example.com --output json webpagetest test https://app.example.com
npm run test:perf # Custom performance suite lighthouse https://app.example.com --output json webpagetest test https://app.example.com

Stress testing

Stress testing

artillery run stress.yml npx loadtest -n 10000 -c 100 https://api.example.com
artillery run stress.yml npx loadtest -n 10000 -c 100 https://api.example.com

Memory and CPU profiling

Memory and CPU profiling

node --inspect script.js python -m cProfile script.py go test -bench=. -benchmem
undefined
node --inspect script.js python -m cProfile script.py go test -bench=. -benchmem
undefined

Output format

输出格式

Performance Test Results:
──────────────────────────────
Load Test (100 concurrent users, 5 minutes):
  ✅ Average Response Time: 245ms (< 500ms target)
  ✅ 95th Percentile: 412ms
  ✅ Throughput: 1,234 req/sec
  ✅ Error Rate: 0.1% (< 1% target)
  ⚠️ CPU Usage: 85% (approaching limit)

Stress Test (Breaking Point):
  ❌ System fails at 850 concurrent users
  ⚠️ Database connection pool exhausted at 800 users
  ✅ Graceful degradation observed

Memory Usage (24-hour endurance):
  ⚠️ Memory leak detected: +2MB/hour
  ❌ OutOfMemory after 18 hours

Summary: Meets most performance targets, needs memory leak fix
性能测试结果：
──────────────────────────────
负载测试（100并发用户，5分钟）：
  ✅ 平均响应时间：245ms (< 500ms 目标)
  ✅ 95分位响应时间：412ms
  ✅ 吞吐量：1,234 次请求/秒
  ✅ 错误率：0.1% (< 1% 目标)
  ⚠️ CPU 使用率：85% (接近上限)

压力测试（断点测试）：
  ❌ 850并发用户时系统故障
  ⚠️ 800并发用户时数据库连接池耗尽
  ✅ 观测到优雅降级表现

内存使用情况（24小时耐久测试）：
  ⚠️ 检测到内存泄漏：每小时增长2MB
  ❌ 运行18小时后出现内存溢出

总结：满足大部分性能指标，需修复内存泄漏问题

Notes

注意事项

  • Establish performance baselines before testing
  • Test in production-like environments
  • Monitor system resources during tests
  • Consider network latency and geography
  • Test with realistic data volumes
  • Automate performance regression testing
  • Document performance requirements and SLAs
  • Use APM (Application Performance Monitoring) tools
  • 测试前先建立性能基线
  • 在类生产环境中开展测试
  • 测试过程中监控系统资源
  • 考虑网络延迟与地域因素
  • 使用符合实际场景的数据量测试
  • 自动化性能回归测试
  • 记录性能需求与SLA
  • 使用APM（应用性能监控）工具