A developer has fixed a bug in a production AWS Lambda function and validated the change in a test environment. The developer wants to roll the update out gradually in production so that initially only 10% of users see the new code. What is the appropriate approach?

The correct approach is to update the Lambda code, publish a new version, and then create a Lambda alias configured for weighted traffic. Lambda aliases allow you to point to a specific function version or split traffic between two versions based on a defined weight. This enables gradual rollouts (canary deployments) where a small percentage of traffic is routed to the new version, allowing for monitoring before a full rollout. Incorrect options: Creating a Lambda trigger to split traffic is not how triggers work; triggers invoke functions based on events, not traffic weighting. Creating a separate Lambda function with the updated code and using an alias to split traffic between two functions is not the standard or most efficient way to handle versioning and gradual rollouts within Lambda. Aliases are designed to manage versions of the same function. A Lambda proxy integration is typically used with API Gateway to pass requests directly to a Lambda function, not for managing weighted traffic between different Lambda versions.