Amazon EC2 Auto Scaling now supports batch instance termination
AWS announced that Amazon EC2 Auto Scaling now allows batch termination of up to 100 instances in a single API call, streamlining scale-down operations for temporary workloads like AI/ML training or container orchestration.
Amazon EC2 Auto Scaling has introduced batch instance termination, enabling users to terminate up to 100 instances in a single API call via the TerminateInstanceInAutoScalingGroup operation. This reduces the number of required API calls when scaling down Auto Scaling groups, particularly for workloads that demand rapid reductions in capacity, such as AI/ML training jobs or container orchestrators. The feature is designed to simplify management for event-driven architectures that temporarily deploy large fleets of instances.
The batch termination process validates all specified instance IDs atomically before proceeding, ensuring consistency and preventing partial execution. Existing Auto Scaling behaviors, including lifecycle hooks and load balancer connection draining, remain intact for each instance in the batch, maintaining operational reliability. This approach minimizes disruptions while allowing precise control over fleet reductions.
The new capability is available across all AWS Regions at no additional cost, aligning with AWS’s standard pricing model for Auto Scaling services. Users can implement batch termination immediately without requiring additional subscriptions or fees, making it accessible for all existing Auto Scaling group configurations.
To implement batch termination, users should refer to the updated Amazon EC2 Auto Scaling User Guide and API Reference Guide for detailed instructions and examples. The documentation outlines the API parameters, validation rules, and best practices for integrating batch termination into existing workflows.