A serverless application uses Amazon API Gateway, AWS Lambda, and an Amazon RDS for PostgreSQL database. During peak or unpredictable traffic, the application experiences database connection timeouts. The company wants to reduce failures with the least code changes. What should a solutions architect do?

Enabling RDS Proxy on the RDS DB instance is the most effective solution because it pools and shares database connections, reducing the overhead of establishing new connections for each Lambda invocation. This prevents connection exhaustion and timeouts during peak traffic without requiring significant code changes. Reducing Lambda concurrency might alleviate the issue but limits scalability and doesn't address the underlying connection management problem. Resizing the RDS instance class can increase the maximum connections but is a more expensive and less efficient solution than connection pooling, and still doesn't prevent connection storms. Migrating to DynamoDB is a significant architectural change, requiring extensive code rewrites, which violates the "least code changes" requirement.