A company runs a critical app using Amazon RDS for MySQL in Multi-AZ. A recent RDS failover test caused a 40-second application outage. A solutions architect needs to reduce outage time to under 20 seconds. Which combination of steps will achieve this? (Choose three.)

Deploying RDS Proxy reduces failover times by maintaining connection pools and automatically routing connections to the new primary instance, masking the failover from the application. Migrating to Amazon Aurora MySQL significantly reduces failover times (typically under 30 seconds, often under 10 seconds) due to its distributed, fault-tolerant storage system and rapid crash recovery. Creating an Amazon Aurora Replica further enhances availability and reduces recovery time objectives (RTOs) because an Aurora Replica can be promoted to primary much faster than an RDS Multi-AZ standby. ElastiCache (Memcached or Redis) helps offload read traffic and improve performance but does not directly reduce the failover time of the underlying RDS database. An RDS for MySQL read replica improves read scalability but does not reduce the failover time for the primary instance; it would also require manual promotion or application changes to use it as a new primary.