SRE Reliability Engineering

Use this skill to design or review production systems through SRE practices. Treat reliability as a managed business and engineering risk: define what users need, measure it with SLI/SLO, budget risk with error budgets, build observability and response loops, and design systems to degrade and recover intentionally.

Core Rule

Start with the user's unresolved reliability decision and the affected journey. Establish the relevant measurement, owner, failure mechanism, action, and recovery evidence. For a narrow task, use only the necessary parts of the broader SRE operating loop.

Hierarchy And Handoffs

This skill owns the operated reliability system after a product or service exists: user-facing SLOs, error budgets, observability, alerting, incident response, on-call, and resilience validation. Use $system-design first when component topology, data flow, service contracts, or capacity choices are still open; bring this skill in to make the selected design measurable and recoverable in production.

Use $software-engineering for codebase and delivery-process health, and $qa-testing for classic pre-release verification. Neither replaces production signals, error-budget policy, or incident ownership.

Use $platform-devops-engineering for infrastructure configuration, apply, and reconciliation. SRE retains runtime diagnosis, user-impact probes, capacity and degradation checks, and operational acceptance. Pass candidate/configuration/state identity, observed behavior, rollback constraints, abort criteria, and owner. Routine red-mainline restoration stays with software/release engineering unless production impact establishes an incident.

For integrity or semantic failures, coordinate with $database-engineering, $data-engineering, or the relevant ML/model owner. They define and repair domain invariants; SRE coordinates containment and verifies operated recovery. Use $application-security-engineering for affected authority and security controls, and $security-review when independent assurance is needed. Availability recovery does not by itself establish restored authorization or data integrity.

Use $technical-writing to shape a runbook, incident report, or operational document. Supply verified behavior, preconditions, state and authority boundaries, failure evidence, version limits, access requirements, and escalation ownership. SRE retains factual ownership and freshness; writing alone needs no additional engineering workflow.

Reference Routing

Read only the references needed for the current task.

• For the topic map and reference index, read references/00_README.md.
• Always start with references/sre-agent-operating-model.md for broad SRE design, review, incident, or production-readiness work.
• For SRE basics, responsibility boundaries, and operating principles, read references/01-sre-foundations.md.
• For SLA/SLO/SLI, CUJ, composite SLO, and SLI implementation choices, read references/02-slo-sli-sla.md.
• For error budgets, burn rate, release policy, and reliability tradeoffs, read references/03-error-budget.md.
• For RED/USE metrics, dashboards, dependency metrics, queues, and whitebox/blackbox monitoring, read references/04-monitoring.md.
• For structured logs, traces, OpenTelemetry, sampling, exemplars, and retention, read references/05-logs-traces.md.
• For page/ticket alerting, multi-window burn-rate alerts, Alertmanager, silence, and signal/noise, read references/06-alerting.md.
• For incident lifecycle, roles, debug docs, planned work, and diagnosis, read references/07-incident-management.md.
• For blameless postmortems, action items, follow-through, and learning loops, read references/08-postmortems.md.
• For on-call rotations, escalation, compensation, handoff, fatigue, and on-call metrics, read references/09-oncall.md.
• For distributed-system failure assumptions, hard/soft dependencies, composite reliability, graceful degradation, and load shedding, read references/10-reliability-architecture.md.
• For integrity incidents, lag and retry feedback, gray failures, independent recovery paths, and restoration of trusted authority, read references/16-recovery-and-integrity.md.
• For Kubernetes resources, runtime limits, probes, graceful shutdown, PDB/HPA, and Helm anti-patterns, read references/11-kubernetes-reliability.md.
• For load/stress testing, chaos experiments, probers, Game Days, and production chaos guardrails, read references/12-chaos-load-testing.md.
• For toil, runbooks, automation, self-service, and automation reliability, read references/13-toil-automation.md.
• For business reliability, culture, local context, burnout, communication, and war-story patterns, read references/14-business-culture-war-stories.md.
• For tools, SLO tooling, observability stack choices, glossary, and current-status guardrails, read references/15-tools-glossary.md.

Workflow

1. Select the reference for the immediate decision using the routing above. Expand to adjacent references only when evidence exposes a dependency; broad readiness work may need several.
2. Identify blocking unknowns. Ask only when the missing data changes the decision; otherwise state assumptions.
3. Tie recommendations to user impact, SLO/error budget, blast radius, owner, validation, rollback, and operational follow-through.
4. Avoid absolute advice where the references describe tradeoffs or maturity prerequisites.
5. Separate observations, hypotheses, derived calculations, and source-backed guidance. State quantitative assumptions and unresolved source defects where they affect the decision.

Output For New System Design

Include:

• Critical user journeys and reliability goals.
• SLI/SLO proposal and assumptions.
• Error budget and release/risk policy.
• Dependency graph with hard/soft dependencies.
• Failure modes, blast radius, and graceful degradation plan.
• Monitoring, logs/traces, dashboard, and alerting plan.
• Rollback, fallback, and mitigation plan.
• On-call, runbooks, incident roles, and postmortem expectations.
• Validation through tests, load/stress, chaos, or drills.
• Ownership, toil/automation plan, risks, and open questions.

Output For Review

Lead with risks and missing decisions:

• Critical production risks.
• Missing or weak SLO/SLI/error-budget decisions.
• Observability and alerting gaps.
• Dependency, timeout, retry, circuit-breaker, fallback, and rollback risks.
• Kubernetes/runtime/probe/shutdown risks when relevant.
• Incident/on-call/postmortem/toil gaps.
• Concrete fixes, validation steps, and owner assumptions.

Output For Incident Support

Prioritize:

1. Safety: avoid worsening the incident.
2. Impact: define affected users, journeys, SLOs, and business effect.
3. Mitigation: rollback, failover, throttling, feature disable, fallback, load shedding.
4. Coordination: IC/Ops/Comms, debug doc, timeline, decision log.
5. Diagnosis: hypotheses by recent changes, dependencies, resources, traffic, data.
6. Recovery: verify user outcomes, protected state, and trustworthy authority before declaring restoration complete.
7. Learning: postmortem and action items after stabilization.

Quality Bar

• Do not call a system reliable without SLOs, observability, actionable alerts, rollback/fallback, and ownership.
• Prefer SLO/burn-rate alerts over raw resource pages unless resource symptoms directly threaten user impact.
• Treat postmortems without owner/deadline/action items as incomplete.
• Treat repeated runbooks and manual fixes as toil candidates.
• Validate the measuring system before using its error budget for policy decisions.
• Name the authority, scope, expiry, and exit condition for exceptions and temporary mitigations.
• Treat numerical thresholds and drill schedules as local decisions supported by workload and risk evidence.
• Require production chaos and load tests to have blast-radius limits, abort conditions, observability, and error-budget headroom.
• When recommending tools, check current project status before presenting them as current best choices.