A company provides an API so customers can retrieve financial information. The company expects spikes in requests during peak periods and requires consistently low-latency responses. The company needs to provide a compute host for the API with the LEAST operational overhead. Which solution should the company use?

The correct solution is to use Amazon API Gateway and AWS Lambda functions with provisioned concurrency. This combination offers the least operational overhead while meeting the requirements. API Gateway handles API management, authentication, and routing without server management. Lambda functions provide serverless compute, scaling automatically. Provisioned concurrency ensures that a pre-initialized number of Lambda instances are ready to respond immediately, eliminating cold starts and guaranteeing consistently low-latency responses during peak periods. Using an Application Load Balancer with Amazon ECS or EKS would involve managing EC2 instances, container orchestration, and scaling policies, leading to higher operational overhead. While API Gateway with reserved concurrency is good, reserved concurrency only prevents other functions from consuming all available concurrency; it doesn't guarantee pre-warmed instances like provisioned concurrency, which is crucial for consistently low latency during spikes.