disaster-recovery
Installation
SKILL.md
Disaster Recovery
Most reliability work assumes the system is basically intact — a pod crashed, a node drained, traffic spiked. Disaster recovery assumes it is not: the region is gone, the primary database is corrupted, someone ran the wrong migration against production. This is a different design problem, and treating it as "just bigger backups" is how organizations discover, mid-disaster, that their backups don't actually restore.
A backup you have never restored is not a backup — it is an unverified belief. The only thing that counts as disaster recovery is a tested path from "everything is gone" to "the business is running again," with a number attached to how long that takes.
A backup you haven't restored is a hope, not a plan.