Amazon RDS and Aurora, particularly for PostgreSQL and MySQL engines, are popular managed relational database solutions in the AWS cloud. Aurora is available in a "serverless" form, but scaling challenges persist due to fixed memory, compute, and network ratios. The article explores the fundamentals of scaling strategies, emphasizing the balance between adding replicas and utilizing caching. Writer replicas handle all writes and require careful management to maintain efficiency, while reader replicas offer eventually consistent reads. Durability and availability are critical, often optimized by using three replicas across three availability zones, which also aligns with AWS's region design. For scaling eventually consistent reads, caching is recommended as a more cost-effective and high-performance solution than adding additional replicas. The emergence of serverless caching solutions like Momento allows for scalable and cost-efficient caching from the outset, providing improved performance and a smoother scaling path for databases like Aurora and RDS. Developers are encouraged to integrate caching early using available frameworks and middleware to enhance application performance and user experience.