Which changes will reduce the lag?

Explanations:

Adding two additional read replicas may help distribute the read load but will not directly address the replication lag issue. The lag is primarily due to the existing read replicas being behind, and simply adding more replicas won’t improve the replication performance unless they are also optimized.

Migrating to an Amazon Aurora MySQL DB cluster can significantly improve replication performance, as Aurora is designed to have lower replication lag and can handle writes more efficiently with a distributed architecture. Aurora also automatically manages the replication process in a more efficient manner compared to standard RDS MySQL.

Moving the read replicas to the same Availability Zone as the primary DB instance might reduce latency slightly but is unlikely to have a significant impact on replication lag if the issue is related to processing capabilities or load. The distance in terms of Availability Zones is typically not the main contributor to lag in this scenario.

Increasing the instance size of the primary DB instance may provide more resources for handling writes, but it won’t directly reduce the replication lag seen on the read replicas, as the lag is primarily due to the read replicas’ processing and their capacity to keep up with the primary instance.

Increasing the instance size of the read replicas to match the primary instance can help reduce replication lag. Larger instances have more CPU and memory resources, allowing them to process replication events more quickly and keep up with the primary DB instance more effectively.