Amazon RDS Multi-AZ (Multi-Availability Zone) is a high-availability feature designed to enhance database durability and fault tolerance. In a Multi-AZ configuration, Amazon RDS automatically maintains a synchronous standby replica of the database in a separate Availability Zone (AZ) within the same AWS Region. This ensures minimal downtime and automatic failover in the event of a primary instance failure.
Primary and Standby Instances:
- The primary database instance handles all read and write operations.
- The standby instance is a synchronous replica maintained in a different AZ for redundancy.
Automatic Failover:
- In case of a failure (e.g., instance crash, AZ outage, or planned maintenance), RDS automatically promotes the standby as the new primary.
DNS Endpoint:
- The application uses a single DNS endpoint, which is automatically updated during a failover, requiring no manual intervention.
Replication:
- Multi-AZ uses synchronous replication, ensuring that data changes on the primary instance are immediately replicated to the standby instance.
| Database Engine | Multi-AZ Supported | Replication Type |
|---|---|---|
| Amazon RDS for MySQL | Yes | Synchronous |
| Amazon RDS for PostgreSQL | Yes | Synchronous |
| Amazon RDS for MariaDB | Yes | Synchronous |
| Amazon RDS for Oracle | Yes | Synchronous (Data Guard) |
| Amazon RDS for SQL Server | Yes | Synchronous (Always On) |
| Amazon Aurora | Natively Multi-AZ | Cluster-based architecture |
Multi-AZ vs Read Replicas
| Feature | Multi-AZ | Read Replicas |
|---|---|---|
| Purpose | High availability and durability | Scaling read workloads |
| Replication Type | Synchronous | Asynchronous |
| Write Operations | Handled by the primary instance | Not supported (read-only) |
| Failover | Automatic | No failover support |
| Cross-Region | Not supported | Supported |
| Use Case | High availability, disaster recovery | Read-heavy applications, analytics |
Multi-AZ Instances Are Not Read Replicas
Multi-AZ (High Availability)
- Synchronous replication: AWS keeps an exact copy of the primary database in another Availability Zone (AZ).
- Automatic failover: If the primary instance fails, AWS promotes the standby instance to minimize downtime.
- Not for read queries: The standby instance cannot be accessed for reads—it only becomes active during failover.
Read Replicas (Read Scaling)
- Asynchronous replication: AWS replicates data from the primary to the read replicas with some delay.
- Can handle read queries: Read replicas can serve read requests, reducing the load on the primary instance.
- Can be promoted: Read replicas can be manually promoted to become a standalone database.