Which feature of Amazon RDS aids in high availability?

Multi-AZ deployments in Amazon RDS are designed specifically to enhance high availability and disaster recovery for database instances. When you configure a database instance with Multi-AZ, Amazon RDS automatically provisions a synchronous standby replica in a different Availability Zone. This standby instance is not accessible for read or write operations under normal circumstances, but it is kept in sync with the primary instance in real-time.

If the primary instance fails due to an issue like a hardware failure, network failure, or maintenance, Amazon RDS automatically performs a failover to the standby instance. This process is managed without any manual intervention, ensuring minimal downtime for applications relying on the database. Thus, applications can quickly resume operations by automatically connecting to the new primary instance, contributing significantly to high availability.

While other features such as Elastic Load Balancing help distribute incoming application traffic, they do not directly influence the availability of the database itself. Data encryption focuses on securing the data at rest and in transit rather than maintaining availability, and connection draining pertains to preventing active requests from being interrupted during maintenance processes, which also does not enhance the core database availability. Therefore, Multi-AZ deployments are the central feature of Amazon RDS that specifically addresses high availability requirements.