AgentSkillsCN

Performance Monitoring

性能监控

SKILL.md

Performance Monitoring

Skill Purpose: Instrument and monitor application performance with metrics, alerts, and trend analysis.


Core Skill Pattern

Objective: Detect performance regressions early and track user experience over time.

Universal Pattern:

  1. Define key metrics (latency, error rate, throughput)
  2. Instrument services and pages
  3. Collect and store metrics
  4. Set alert thresholds and dashboards
  5. Review trends and regressions

Key Decisions (Project-Specific):

  • Monitoring tooling and data retention
  • Alert thresholds and escalation rules
  • Sampling rates for client metrics

Project-Specific Implementation Notes

Customize per project:

  • Include Core Web Vitals for web apps
  • Distinguish backend vs frontend metrics

Example Metrics

  • p50/p95 response time
  • error rate by endpoint
  • Core Web Vitals
  • bundle size trends

Best Practices

  1. Alert on user-impacting regressions only
  2. Use percentiles, not averages
  3. Keep dashboards focused and actionable
  4. Review performance weekly
  5. Tie alerts to runbooks

Stop Conditions

STOP and escalate if:

  • No metrics are defined
  • Alerts are noisy or missing
  • Monitoring does not cover critical paths

Skill Version: 1.0.0