sre-engineer_skill

Overview

This skill encapsulates senior Site Reliability Engineering practices for defining SLIs/SLOs, managing error budgets, and building reliable systems at scale. It focuses on measurable outcomes: availability targets, error-budget policies, and automation to reduce toil while preserving developer velocity. Use it for incident response, chaos engineering, capacity planning, and on-call improvements.

How this skill works

I assess architecture, current SLOs, incident history, and toil to produce quantitative reliability plans. I define actionable SLIs, set SLO targets, calculate error budgets and burn rates, and produce monitoring, alerting, and automation artifacts (dashboards, Prometheus rules, runbooks, scripts). I also design chaos experiments and capacity models and provide blameless postmortems and remediation steps.

When to use it

• When defining or revising SLIs, SLOs, or error-budget policies
• When incidents recur or burn rates threaten feature velocity
• To reduce operational toil via automation
• To design and validate resilience with chaos engineering
• For capacity planning and on-call practice improvements

Best practices

• Always justify SLO targets with user-impact data and business context
• Measure golden signals (latency, traffic, errors, saturation) continuously
• Calculate error budgets from SLOs and treat exhaustion as a policy trigger
• Automate repetitive tasks and track toil reduction metrics
• Write blameless postmortems with corrective actions and owners

Example use cases

• Define a 99.9% availability SLO for a public API, with Prometheus alerts and an error-budget policy
• Create a runbook and automation to reduce a common manual failover task
• Design a chaos experiment that validates graceful degradation under instance failures
• Build a capacity plan based on traffic forecasts and scaling characteristics
• Investigate repeated incidents, produce a blameless postmortem, and propose fixes