A Java application was deployed to AWS Elastic Beanstalk using a single-instance environment to save costs during development. Testing shows sustained CPU utilization above 85%, causing performance bottlenecks. You must mitigate the CPU bottleneck before production with the least operational overhead. What should you do?

The correct option directly addresses the problem with the least operational overhead by modifying the existing environment. Changing the capacity configuration to a load-balanced environment type allows for horizontal scaling, distributing the load across multiple instances and mitigating the CPU bottleneck. Enabling all Availability Zones improves fault tolerance and availability. Adding a scale-out rule based on average CPU utilization ensures that new instances are launched automatically when demand increases, maintaining performance. Creating a new Elastic Beanstalk application (first incorrect option) would require redeploying the application and reconfiguring settings, increasing operational overhead. Creating a second environment with traffic splitting (second incorrect option) is a deployment strategy, not a direct solution for scaling a single overloaded environment. The "Rebuild environment" action (fourth incorrect option) is for recreating the environment from scratch, which is more disruptive than modifying the existing configuration and doesn't inherently solve the scaling problem without further configuration. Additionally, it specifies "sum CPU utilization," which is less appropriate for scaling than "average CPU utilization.