When something customer-facing breaks, the goal is to restore service first and understand it second.

Severity levels

During an incident

1. Declare it in the incidents channel and name an incident lead.
2. Mitigate: roll back using the Deployment runbook or disable the feature flag.
3. Update the status page every 30 minutes until resolved.
4. Write the review within five working days.

Blameless does not mean nameless. We name the systems that failed, not the people who used them.