⚡AgentSkills
☁️ DevOps & Cloud · Observability

Instrument the four golden signals

Latency, traffic, errors, saturation per service with SLO thresholds that map to user pain.

intermediate~35 minDevOps EngineersSREsCloud Architects

Steps

  1. 1Emit RED metrics per endpoint: rate, errors, duration histograms
  2. 2Track saturation of the true bottleneck (connections, queue depth, CPU steal)
  3. 3Define SLOs from user journeys, not server uptime vanity
  4. 4Burn-rate alerts on fast+slow windows to catch real incidents early
  5. 5Dashboards ordered by journey, not by org chart
  6. 6Review signal usefulness quarterly; delete dashboards nobody opens

Common Pitfalls

  • ▲CPU alerts while users suffer on queue lag
  • ▲Averages hiding p99 cliffs

Commands

Install with skills CLI
$ npx skills add aniruddhaadak80/skills --skill observability-golden-signals
Install globally
$ npx skills add aniruddhaadak80/skills --skill observability-golden-signals -g

Tags

#monitoring#metrics#slo#devops-cloud#observability

Related skills

Adopt structured logging that answers questionsfoundation

JSON logs with trace correlation and disciplined fields you can actually query at 3am.

☁️ DevOps & Cloud·~25m
Open →