Skip to main content

Database DR and Replication Options

RDS Disaster Recovery / Replication Options


a) Multi-AZ Deployment (same region)
  • What it does: AWS runs a standby instance in a different AZ within the same region. Storage is synchronous.
  • Failover: Automatic within minutes if the primary fails.
  • Cost impact: You pay for the standby instance as if it’s another RDS, so ~100% of your current RDS instance cost again.
  • Use case: Improves uptime, but doesn’t protect from full region failure.

b) Read Replica in Another Region (Cross-Region Replication)
  • What it does: Asynchronous replication to a DB in another AWS region.
  • Failover: Manual promotion required if the primary fails.
  • Cost impact: You pay full price for the replica instance + cross-region data transfer (usually a few % of RDS cost for moderate workloads). (+ ~100% + $10-20 Data Transfer)
  • Use case: Protects against full region outage, but not instant failover.

c) Multi-AZ Primary + Cross-Region Replica
  • What it does: Combines both — the primary DB has a same-region standby and a read replica in another region.
  • Failover:
    • Within same region: automatic to standby.
    • Region-wide outage: promote replica manually.
  • Cost impact:
    • Standby instance (~100% of RDS cost).
    • Replica (~100% of RDS cost) + data transfer fees.
    • Essentially ~3× your current RDS compute cost, plus some transfer/storage overhead. + data transfer

d) Periodic Snapshot & Restore Cross-Region (This needs to be done some or the other way for backups)
  • What it does: Takes DB snapshots in the primary region and copies them to another region on a schedule (daily/weekly).
  • Failover: Restore snapshot in new region — downtime can be hours.
  • Cost impact: Snapshot storage (~$0.095/GB-month) + transfer cost; cheapest option but slow recovery.
  • Use case: Low-cost compliance-friendly DR; good if RTO can be long.