A website backed by an Amazon Aurora cluster is experiencing slow response times during peak periods. Users perform searches that are seldom repeated, and traffic patterns vary by season and day. To improve performance while using resources efficiently, which solution should the administrator implement?

Creating Aurora Replicas and using the reader endpoint offloads read traffic, like search queries, from the primary instance, improving response times. Aurora Auto Scaling on replicas automatically adjusts read capacity based on demand, ensuring efficient resource utilization during varying traffic patterns. This approach specifically addresses read-heavy workloads and fluctuating demand. Placing an ElastiCache for Redis cluster would be effective for frequently repeated searches, but the problem states searches are "seldom repeated," making caching less impactful. Provisioning IOPS on the storage layer might help with I/O bottlenecks but doesn't distribute the read load. Increasing the DB instance class would handle peak traffic but is less cost-effective for varying demand and doesn't distribute read load as efficiently as replicas.