High Availability & Disaster Recovery
Designing for component failure as the expected case, and the RTO/RPO trade-off that shapes disaster-recovery strategy.
Availability as a design goal
High availability means the system keeps working despite individual component failures — achieved through redundancy (multiple instances, no single point of failure) spread across failure domains (multiple availability zones, ideally multiple regions), plus health checks that detect and route around a failed instance automatically.
RTO and RPO
Two numbers define a disaster-recovery strategy: Recovery Time Objective — how long can the system be down before it must be back up — and Recovery Point Objective — how much data can acceptably be lost (the gap since the last backup/replication checkpoint). Tighter RTO/RPO requires more infrastructure investment (real-time replication, automated failover) versus looser targets that tolerate periodic backups and manual recovery.
Common DR strategies, by cost and speed
- Backup & restore — cheapest, slowest recovery (hours+).
- Pilot light — minimal standby infrastructure, scaled up on failover (tens of minutes).
- Warm standby — a smaller-scale but running replica, promoted on failover (minutes).
- Multi-site active-active — full redundant capacity running in parallel at all times, near-instant failover, most expensive.
Prerequisite
Combines IAM (securing the redundant infrastructure identically) and Infrastructure as Code (reliably reproducing environments for failover) — which is why both come first.