A company created an API in Amazon API Gateway that invokes Lambda functions and deployed it to a Production stage. An external development team is the only API consumer. Usage spikes at certain times, raising cost concerns. The company wants to limit usage and costs without modifying the Lambda functions. What is the MOST cost‑effective solution?

The most cost-effective solution is to create an API Gateway API key and a usage plan with throttling limits and quotas. This directly addresses the requirement to limit usage and costs without modifying Lambda functions. The usage plan allows you to define specific request rates and burst limits, as well as total request quotas over a period, which directly controls the number of invocations and thus the cost. Sharing the API key with the external team ensures they are subject to these defined limits. Incorrect options: Sending requests to SQS and then polling with Lambda adds complexity and potential latency, and doesn't inherently limit the number of requests processed by Lambda, only how they are invoked. Provisioned concurrency for Lambda is for performance optimization, not cost reduction from usage limits, and Application Auto Scaling would increase, not decrease, costs during spikes. An AWS WAF Regional web ACL with a rate-based rule can limit requests, but an API Gateway usage plan is a more direct and integrated feature for managing API consumption and associated costs, especially when dealing with specific consumers and quotas.