AWS Lambda Increases Timeout to 90 Minutes for Managed Instances

The increase in timeout allows AWS Lambda to handle longer-running workloads, which can simplify deployments and reduce the need for workarounds. This change can be particularly beneficial for use cases such as media processing, financial calculations, and data processing, where jobs can exceed 15 minutes of continuous execution.

However, the increased timeout also requires extra care with network connections, temporary credentials, and duplicate execution. Developers need to ensure that connections and credentials remain valid for the full runtime and design operations to safely handle retries and duplicate processing. This can be achieved by using tools such as Powertools for AWS Lambda to implement idempotency in function code.

The community reaction to the increased timeout has been mostly positive, but some warn that it might encourage fragile and costly patterns. It is essential for developers to carefully consider the economics of using AWS Lambda for long-running workloads and to evaluate alternative options, such as ECS Tasks or AWS Batch, if necessary. The increased timeout is supported in all regions where Lambda Managed Instances are available, and it is also supported by Graviton5-powered EC2 instances.