SLO Targets, Budgets, and Burn-Rate Alerts
Get measurable SLIs, justified SLO targets, correct error budgets, and paired burn-rate alerts with an operating policy.
Before & after
Same prompt, same agent — the only difference is the skill.
Service level objectives help you define reliability targets for your systems. You typically set a percentage uptime or availability goal. Common targets are 99.9percent or 99.95percent availability. You'd also want to think about different types of failures - some might be more critical than others. Error budgets can help you decide when it's safe to take risks like major refactors or deploy without waiting. You'd need to define what actually counts as a success or failure for your service. Monitoring and alerting are important to track whether you're meeting your targets.
SLO DEFINITION FOR PAYMENT PROCESSING: SLI METRIC DEFINITION (Service Level Indicator) Measurement: Successful payment transactions / Total payment attempts Calculation: count(payment.status = 'settled') / count(payment.status IN 'attempted', 'settled', 'failed') Exclusions: Test transactions (env='test'), Manually refunded (category='refund') Window: 30-day calendar month SLO TARGET: 99.5percent - Equals 3.6 hours downtime/month - Rationale: Payment is critical path; 99.9percent (43 min/month) overly stringent for infrastructure variance ERROR BUDGET ALLOCATION (43.2 minutes/month available) Planned maintenance: 15 minutes (database backup window 2am-2:15am weekdays) Infrastructure incidents: 20 minutes (DNS failover, load balancer reboot contingency) Deployment risk: 8.2 minutes (reserved for emergency rollbacks) Burn rate thresholds: • >14.4 min/week = YELLOW (stop non-critical changes, page team) • >7.2 min/week = RED (halt all deploys, incident response) MULTI-WINDOW BURN RATE ALERTS 1-hour window: if error rate 10X baseline → page on-call 6-hour window: if error rate 5X baseline → alert team lead 24-hour window: if budget burn >20percent monthly → meeting required SLO REVIEW MONTHLY Track: Were SLI targets met? What caused misses? Adjust: Next month's error budget spend based on learnings Communicate: Share brief 'how we did' report with stakeholders
About this skill
“Our SLO is green, but customers are still failing checkout.” “The burn-rate alert pages constantly, or fires after the budget is gone.” Turn reliability targets into explicit measurements, correct budget calculations, and alerts tied to an action someone can take.
What it does
- Defines exactly which events count, which failures are bad, where measurement happens, and how latency is evaluated.
- Selects provisional or evidence-backed targets using user tolerance, historical performance, and dependency constraints.
- Calculates request-based or time-based budgets without confusing failed requests with downtime minutes.
- Builds paired short- and long-window burn-rate conditions with thresholds derived from budget consumption.
- Handles low traffic, absent telemetry, changing traffic volume, and aggregate metrics that hide regional failures.
- Sets budget-based release decisions, exception ownership, and recovery criteria.
What you get
An implementable SLO specification, transparent budget arithmetic, a burn-rate alert plan, and an operating policy. For an existing proposal, get specific defects and corrected definitions or expressions; quick questions receive focused answers.
Best for
SREs, platform engineers, backend service owners, and engineering leads defining their first SLO or making an existing reliability program operational.
SKILL.md preview
---
name: slo-targets-budgets-and-burn-rate-alerts
description: 'Define and operate user-facing SLIs, evidence-based SLO targets, event or time error budgets, and paired burn-rate alerts. Derive alert thresholds from budget consumption and connect them to an explicit operating policy. Use when the user says "define an SLO", "what should our reliability target be", "calculate our error budget", "our burn-rate alerts keep paging", "customers are failing but our SLO is green", "review this SLI", or "what happens when the budget runs out".'
---
# SLO Targets, Budgets, and Burn-Rate Alerts
… (sign up to view the full skill)More development skills
View all Development skills →Chaos Experiment and Game Day Planner
Plans chaos experiments and game days: hypothesis, blast radius, abort conditions, tool manifests, rollback, and post-run review.
Claude Code Cost and Latency Routing
Configure Claude Code Router with model tiers, explicit fallbacks, compatibility checks, and a cost-aware validation plan.
CI/CD Pipeline Builder with Gated Deploys
Generates CI/CD workflows from your repo's real stack: cached fast-fail checks, pinned actions, and staging-to-prod deploys with gates.
Terraform Modules and Safe Deployments
Get reusable Terraform modules, safe state migrations, explicit provider wiring, and CI gates tied to reviewed plans.