A website displays a daily newsletter stored in English. When a user visits, a Lambda function queries the company's on-premises database for the current newsletter, calls Amazon Translate TranslateText to translate it for the user, and returns the translated text. Traffic has increased and the on-premises database is overloaded, slowing responses. The company cannot change the database. Which change will improve the Lambda function's response time?

Caching the translated newsletters in the Lambda function's /tmp directory (or an external cache like ElastiCache or DynamoDB) is the most effective solution. This reduces the number of calls to the overloaded on-premises database and Amazon Translate, directly addressing the bottlenecks. The /tmp directory is suitable for temporary storage within a Lambda execution environment. Switching to asynchronous invocations would not improve response time for the user; it would only change how the Lambda function is invoked. Enabling caching for TranslateText API calls is not a standard feature; Translate is a real-time translation service. Modifying the Lambda function to process requests in parallel would further stress the overloaded on-premises database, worsening the problem.