Light Production Observability

Use this skill when a system needs to be debuggable in production without guesswork. It is intentionally broader than dashboards: the output should make an operator able to answer "what failed, where, why, for whom, and how do we know the fix worked?"

For multi-step provisioning, ingestion, migration, sync, CI/CD, or external API orchestration, use observable-workflow-design instead of relying on this lighter checklist alone. For formal design reviews, PR reviews, or launch gates, prefer production-observability.

Default Stance