SLO Architect
SLO Architect is a skill for defining, reviewing, and operating Service Level Objectives (SLOs), SLIs, and error budgets. It provides three Python tools to design SLOs, calculate error budgets with multi-window burn-rate alerts, and review existing SLOs for common bugs, following Google SRE Workbook principles.
✨ What it does
- Generates structured SLO definitions with required fields
- Calculates error budgets and multi-window burn-rate alert thresholds
- Reviews SLOs for common bugs like target too high, wrong SLI, no error budget policy
- Supports multiple SLI types: request success rate, latency, availability, data freshness, correctness
- Provides composition with feature-flags-architect, chaos-engineering, and kubernetes-operator
🎯 When to use it
- Defining a new SLO for a service or feature
- Reviewing existing SLOs for common bugs
- Computing error budgets and burn-rate alert thresholds
- Tying SLOs to existing controls like feature flags or chaos engineering
🚀 How to use
Trigger: /slo-design
Trigger by asking to define an SLO, compute error budgets, or review existing SLOs. Provide service name, SLI type, target availability, and window. Use the slash command /slo-design for an interactive wizard. Example prompts:
Define an SLO for checkout-svc with 99.9% request success rate over 30 days.
Calculate error budget for 99.95% availability over 7 days.
Review SLOs in docs/slos/ for common bugs.
📄 Output: The skill produces SLO definition files, error budget calculations with alert rules, and review reports identifying issues.
📦 Add this skill to Claude Code
# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills
# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/engineering/skills/slo-architect .claude/skills/slo-architectSkill source: engineering/skills/slo-architect/SKILL.md
⚠️ Good to know
Requires Python 3 and is not for general observability, SLA drafting, load testing, or incident response.
❓ FAQ
What SLI types are supported?
The skill supports request-success-rate, request-latency, availability-time, data-freshness, and correctness.
How does the skill determine if an SLO target is too high?
The slo_review.py tool flags targets ≥99.99% as too high, as they are sustainable only with massive engineering investment.
What is the recommended SLO window?
The skill recommends a window of 28 days (4 calendar weeks) for SLOs, and flags windows shorter than 7 days or longer than 90 days as problematic.
🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.