A web application uses Amazon API Gateway, AWS Lambda, and Amazon DynamoDB. After a marketing campaign, many requests have much longer response times. CloudWatch Logs for API Gateway show errors on 20% of requests. For the Lambda function, the Throttles metric is 1% and Errors is 10%. Application logs show that errors coincide with calls to DynamoDB. What change should the solutions architect implement to improve response times as traffic grows?

The problem indicates that errors coincide with calls to DynamoDB, suggesting a bottleneck there. DynamoDB’s ThrottledRequests metric would confirm this. Enabling DynamoDB auto scaling allows the table to automatically adjust its read and write capacity units based on actual traffic patterns, preventing throttling and improving response times during traffic spikes. Increasing the Lambda function's concurrency limit is not the primary solution because the Lambda Throttles metric is only 1%, and the errors are tied to DynamoDB calls, not Lambda concurrency. Increasing API Gateway's throttle limits would allow more requests to reach the backend, but it wouldn't resolve the underlying DynamoDB bottleneck, potentially worsening the problem. Re-creating the DynamoDB table with a better primary key is a more complex and disruptive change that might be necessary for long-term optimization, but auto scaling is the immediate and most effective solution for handling increased traffic and preventing throttling errors.