operability
Operations - deployment, rollback, feature flags. Use for ops tooling.
When & Why to Use This Skill
The Operability skill for Claude is designed to bridge the gap between software development and production stability. It focuses on SRE (Site Reliability Engineering) principles, enabling teams to manage deployments, rollbacks, and feature flags with a focus on system resilience. By emphasizing 'operability,' this skill ensures that production systems are not just built, but are maintainable, providing tools for incident remediation, dead-letter handling, and drift detection to resolve issues without emergency code deployments.
Use Cases
- Incident Remediation: Providing operators with actionable runbooks and tools to diagnose and fix stuck jobs or failed workflows during production outages.
- Safe Deployment & Rollback: Managing high-stakes releases using feature flags and executing rapid rollbacks when critical system drift or errors are detected.
- Workflow Recovery: Safely replaying failed background tasks (e.g., via Upstash) by ensuring side-effects like billing or emails are idempotent and won't be duplicated.
- Drift Management: Monitoring for inconsistencies between system states and applying automated remediation playbooks to maintain infrastructure integrity.
| name | operability |
|---|---|
| description | Operations - deployment, rollback, feature flags. Use for ops tooling. |
Operability Guideline
Tech Stack
- Workflows: Upstash Workflows + QStash
- Cache: Upstash Redis
- Platform: Vercel
Non-Negotiables
- Dead-letter handling must exist and be operable (visible, replayable)
- Side-effects (email, billing, ledger) must be idempotent or safely re-entrant
- Drift alerts must have remediation playbooks
Context
Operability is about running the system in production — not just building it. Systems fail. Jobs get stuck. State drifts. The question is: when something goes wrong, can an operator fix it without deploying code?
Consider the operator experience during an incident. What tools do they have? What runbooks exist? Can they safely retry failed jobs? Can they detect and fix drift?
Driving Questions
- What happens when a job fails permanently?
- How would an operator know something is stuck?
- Can failed workflows be safely replayed without duplicating side-effects?
- What drift can occur between systems, and how would we detect it?
- What's the rollback plan if a deploy breaks something critical?
- What runbooks exist, and what runbooks should exist but don't?