Application developers have reported that an application is running slower as more users are added. The application database is running on an Amazon Aurora DB cluster with an Aurora Replica. The application is written to take advantage of read scaling through reader endpoints. A database specialist looks at the performance metrics of the database and determines that, as new users were added to the database, the primary instance CPU utilization steadily increased while the Aurora Replica CPU utilization remained steady.
How can the database specialist improve database performance while ensuring minimal downtime?

A. Modify the Aurora DB cluster to add more replicas until the overall load stabilizes. Then, reduce the number of replicas once the application meets service level objectives.
B. Modify the primary instance to a larger instance size that offers more CPU capacity.
C. Modify a replica to a larger instance size that has more CPU capacity. Then, promote the modified replica.
D. Restore the Aurora DB cluster to one that has an instance size with more CPU capacity. Then, swap the names of the old and new DB clusters.