Disaster recovery

Disaster recovery (DR) prepares for region loss, datacenter failure, ransomware, or operator error destroying primary infrastructure. Plans include backup storage in separate accounts/regions, runbooks for failover, regular game days, communication templates, and defined roles during incidents.

Active-passive DR is simpler; active-active reduces RTO but adds consistency complexity. Document data loss window (RPO) and time to restore service (RTO) with realistic drill results, not aspirational slides.

On interviews: outline DR for a single-region MVP evolving to multi-region, failover steps, and what data you might lose in worst case.

Common pitfalls: DR backups in same region as primary; runbooks untested for years; DNS failover without TTL planning; no customer communication plan.

The trade-off is flexibility versus complexity—know when the simpler path is enough.

Checklist:

• Define RTO and RPO with stakeholders.
• Store backups in isolated region/account.
• Run failover drills on a schedule.
• Maintain runbooks and comms templates.