AWS Application Load Balancer launches Target Optimizer
Application Load Balancer (ALB) now offers Target Optimizer, a new feature that allows you to enforce a maximum number of concurrent requests on a target.
With Target Optimizer, you can fine-tune your application stack so that targets receive only the number of requests they can process, achieving higher request success rate, more target utilization, and lower latency. This is particularly useful for compute-intensive workloads. For example, if you have applications that perform complex data processing or inference, you can configure each target to receive as few as one request at a time, ensuring the number of concurrent requests is in line with the target's processing capabilities.
You can enable this feature by creating a new target group with a target control port. Once enabled, the feature works with the help of an agent provided by AWS that you run on your targets that tracks request concurrency. For deployments that include multiple target groups per ALB, you have the flexibility to configure this capability for each target group individually.
You can enable Target Optimizer through the AWS Management Console, AWS CLI, AWS SDKs, and AWS APIs. ALB Target Optimizer is available in all AWS Commercial Regions, AWS GovCloud (US) Regions, and AWS China Regions. Traffic to target groups that enable Target Optimizer generates more LCU usage than regular target groups. For more information, see the pricing page, launch blog, and ALB User Guide.