Amazon ECS Managed Instances (ECS Managed Instances) now gives you greater control over infrastructure optimization with configurable scale-in delay. This enhancement allows you to fine-tune instance management based on your specific workload patterns and business requirements, helping you better balance cost optimization with operational needs.
ECS Managed Instances is a fully managed compute option that automatically provisions right-sized Amazon EC2 instances based on your workload requirements. Over time, your compute resources may drift from workload requirements due to changing traffic patterns or dynamic scaling. ECS Managed Instances continuously monitors and proactively optimizes costs by terminating idle Amazon EC2 instances not running any tasks, and consolidating tasks from underutilized instances onto other, right-sized instances, provisioning new instances if required. ECS uses a heuristic based delay for scaling-in your instances to deliver a balance of high availability and cost optimization. However, your workloads or business may have unique requirements. For example, you might need to retain instances for a longer time period to accommodate incoming batch jobs and minimize instance churn. Starting today, you can set the scaleInAfter configuration parameter to up to 60 minutes to align with your specific infrastructure optimization needs. You can also set the scaleInAfter to -1 to disable infrastructure optimization workflows, which will allow your instances to run until they are patched after 14 days.
You can use ECS API, console, SDK, CDK, CloudFormation to configure scaleInAfter parameter when creating or updating an ECS Managed Instances capacity provider. This feature is available in all commercial AWS Regions. To learn more, review documentation and deep dive blog post.
Categories: general:products/amazon-ecs,marketing:marchitecture/compute,marketing:marchitecture/containers
Source: Amazon Web Services
