advanced

RTO / RPO

State recovery time and recovery point objectives so design cost matches downtime and data-loss tolerance.

RPO (Recovery Point Objective) is maximum acceptable data loss measured in time — how far back backups replay. RTO (Recovery Time Objective) is maximum acceptable downtime until service restores. Tighter objectives cost more: synchronous replication, multi-region active-active, and frequent snapshots.

Align objectives with business impact: email delay tolerates minutes of RPO; payment ledger may require seconds. Engineering proposes options with cost; product owns the trade-off.

On interviews: given a prompt, propose RTO/RPO for core flows and list infrastructure choices that satisfy them.

Common pitfalls: conflating backup frequency with RPO without testing restore; RTO ignoring DNS and cache warmup; identical objectives for all data tiers.

Checklist:

State RPO and RTO per critical flow.
Map objectives to replication and backup design.
Validate with restore drill metrics.
Tier objectives by business impact.