High Availability & DR

HA & DR Quick Notes

Definitions

• RPO (Recovery Point Objective): How much data loss is acceptable (time).
• RTO (Recovery Time Objective): How long downtime is acceptable.

In-Region HA Patterns

• Multi-AZ load balancers + ASG across AZs.
• RDS / Aurora Multi-AZ for synchronous failover.
• NAT Gateway per AZ for private egress HA.

DR Strategy Ladder (cold → hot)

Strategy, Idea, RTO/RPO feel

Backup & Restore: Snapshots / backups only, Highest RTO

Pilot Light: Minimal core always on in AWS, Medium

Warm Standby: Scaled-down but always-on copy, Lower

Multi-site active/active: Full capacity in ≥2 sites, Lowest

Cross-Region Patterns

• Aurora Global Database (RPO seconds / RTO ~minute class).
• DynamoDB global tables.
• S3 CRR.
• Route 53 failover + warm stack in AWS (e.g. Storage Gateway stored volumes already syncing to S3).

Tips

• US-only users → prefer multi-AZ in one Region over multi-Region complexity.
• “Least downtime” hybrid failover → warm resources already running + DNS failover — not “launch CloudFormation only when disaster hits.”