返回 Skill 列表
extension
分类: 开发与工程无需 API Key

performance-test-planning

设计性能测试策略,包括负载概况、容量规划、SLA验证以及使用NBomber和k6进行.NET性能测试。

person作者: jakexiaohubgithub

Performance Test Planning

When to Use This Skill

Use this skill when:

  • Performance Test Planning tasks - Designing load profiles, capacity planning, SLA validation
  • Planning or design - Need guidance on performance testing approaches
  • Best practices - Want to follow established patterns and standards

Overview

Performance testing validates that systems meet non-functional requirements under expected and peak loads. Effective planning identifies critical paths, defines realistic load profiles, and establishes measurable success criteria.


Performance Test Types

| Type | Purpose | Load Pattern | Duration | |------|---------|--------------|----------| | Load Test | Validate expected load | Normal traffic | 15-60 min | | Stress Test | Find breaking point | Increasing until failure | Until failure | | Soak Test | Find memory leaks | Sustained normal load | 4-24 hours | | Spike Test | Handle sudden bursts | Sharp increase/decrease | 10-30 min | | Capacity Test | Determine max capacity | Incremental increase | Variable | | Scalability Test | Validate horizontal scale | Increasing with resources | Variable |


Load Profile Patterns

Ramp-Up Pattern

Users
  │
100├────────────────────────●
  │                   ●────
  │              ●────
50├         ●────
  │    ●────
  │●────
 0└─────────────────────────►
   0    1    2    3    4    5 min

Spike Pattern

Users
  │         ●
  │        ╱ ╲
500├──────╱   ╲──────
  │     ╱     ╲
  │    ╱       ╲
100├───●         ●───
  │
 0└─────────────────────────►
   0   5   10  15  20 min

Quick Reference: Metrics Targets

| Metric | Target | Critical Threshold | |--------|--------|-------------------| | Response Time (p50) | < 100ms | < 200ms | | Response Time (p95) | < 200ms | < 500ms | | Response Time (p99) | < 500ms | < 1000ms | | Throughput | > 1000 RPS | > 500 RPS | | Error Rate | < 0.1% | < 1% | | CPU Utilization | < 70% | < 90% |


Key Formulas

Little's Law: L = λ × W (Concurrent requests = Arrival rate × Response time)

Throughput: Concurrent Users / Average Response Time

Apdex: (Satisfied + Tolerating/2) / Total


References

| Reference | Content | When to Load | | --- | --- | --- | | performance-strategy-template.md | Full strategy template, scope, environment, scenarios | Planning performance test strategy | | nbomber-examples.md | Basic load test, complex scenarios, data-driven tests | Implementing .NET performance tests | | capacity-planning.md | Little's Law, throughput, Apdex, BenchmarkDotNet | Capacity planning calculations |


Integration Points

Inputs from:

  • Non-functional requirements → SLA targets
  • Architecture documents → Component topology
  • test-strategy-planning skill → Test scope

Outputs to:

  • CI/CD pipeline → Automated performance gates
  • Monitoring systems → Baseline metrics
  • Capacity planning → Infrastructure sizing

Test Scenarios

Scenario 1: Planning performance strategy

Query: "Help me create a performance test plan for our API"

Expected: Skill activates, provides strategy template, guides through objectives and scope

Scenario 2: Implementing load tests

Query: "Show me how to use NBomber for load testing in .NET"

Expected: Skill activates, loads nbomber-examples.md reference, provides code examples

Scenario 3: Capacity planning

Query: "How do I calculate capacity requirements for 10,000 users?"

Expected: Skill activates, loads capacity-planning.md reference, provides formulas and templates


Last Updated: 2025-12-28

Version History

  • v1.1.0 (2025-12-28): Refactored to progressive disclosure - extracted templates/examples to references/
  • v1.0.0 (2025-12-26): Initial release