Alloydb Continuous Backup Pitr Strategy

What Continuous Backup and Recovery Is

AlloyDB continuous backup and recovery is a native data protection mechanism that continuously records incremental storage changes and transaction logs. This capability allows administrators to restore an existing cluster to any precise moment in its recent history with microsecond granularity, eliminating data loss gaps between scheduled backup intervals.

How It Meets the Requirements

• Near-Zero RPO: By continuously logging database state changes, continuous backup enables a Recovery Point Objective (RPO) of zero, allowing restoration up to the exact moment right before a corruption event or accidental deletion.
• Specific Timestamp Restoration: Unlike discrete snapshot schedules that only capture periodic states, Point-in-Time Recovery (PITR) provides granular rewind capability to any selected second.
• Two-Week Window: By default, AlloyDB retains continuous backup data for 14 days (configurable from 1 to 35 days), directly matching the two-week retention mandate.

Key Advantages and Capabilities

• Incremental efficiency: Stores only changed blocks relative to previous backups, minimizing storage overhead and cost.
• Automated daily baseline: Automatically captures daily incremental backups alongside continuous write-ahead log data.
• Cluster cloning: Enables rapid creation of independent clones from any timestamp within the retention window without altering the source cluster.
• Immutable retention: Ensures underlying backup files are protected against tampering or unauthorized modification.

Why This Is Optimal

Continuous backup and recovery is the native managed solution specifically designed for relational database point-in-time recovery. It achieves an RPO of zero and provides microsecond-level recovery points without requiring custom scheduling logic, external orchestration, or disruptive full-database lock operations.