The correct answer is Enable Multi-AZ configuration for the Amazon RDS instance to ensure synchronous data replication across Availability Zones. Multi-AZ deployments in Amazon RDS use synchronous replication to a standby instance in a separate Availability Zone, which guarantees minimal data loss and supports an RPO of less than 1 second. This setup is designed specifically for high availability and disaster recovery, making it ideal for production databases with strict compliance or operational standards.

The option Run the RDS instance in a single Availability Zone and use AWS Database Migration Service (AWS DMS) with change data capture (CDC) enabled is incorrect because DMS is intended for migration and replication scenarios. While CDC can reduce lag, it is asynchronous and not designed to meet sub-second RPOs in production environments.

The choice Deploy the RDS instance in one Availability Zone and add multiple read replicas in another zone for failover support is unsuitable because read replicas are also asynchronously replicated. This means they cannot guarantee an RPO of under 1 second and are more appropriate for read scaling than for high-availability failover.

The option Configure auto scaling for the RDS instance to adjust capacity in response to usage spikes is unrelated to RPO requirements. Auto scaling addresses performance under load, not recovery from failure or data protection.

This question is important because it distinguishes between high availability, disaster recovery, and performance tuning. To answer it correctly, focus on keywords like "RPO less than 1 second," which require synchronous replication. Only Multi-AZ meets this condition in RDS for PostgreSQL.

The challenge here is avoiding plausible but incorrect options like read replicas or CDC, which are useful in other scenarios but not when ultra-low RPO guarantees are required. Understanding the difference between asynchronous and synchronous replication is critical.

Full AWS Practitioner Certification Question