A company runs an application on AWS that receives variable amounts of traffic. The application uses AWS Direct Connect to an on-premises MySQL-compatible database. The on-premises database consistently uses a minimum of 2 GiB of memory. The company wants to migrate the database to a managed AWS service and use auto scaling to handle unexpected increases in workload. Which solution meets these requirements with the LEAST administrative overhead?

Aurora Serverless v2 is the best solution because it offers true on-demand scaling, automatically adjusting capacity based on workload with fine-grained increments, minimizing administrative overhead. Setting a minimum capacity of 1 ACU ensures it meets the 2 GiB memory requirement (1 ACU provides 2 GiB of memory). Amazon Aurora provisioned (Option B) scales, but not as elastically or automatically as Serverless v2, requiring more manual intervention for scaling events. Amazon RDS for MySQL (Option D) also requires manual scaling or more complex auto-scaling configurations, increasing administrative overhead. Amazon DynamoDB (Option A) is a NoSQL database, which would require significant application re-architecture to migrate from a MySQL-compatible database, and it doesn't directly map to the 2 GiB memory requirement in the same way.