AWS Auto Scaling Group Stuck in 'Updating Capacity' During Scale-In

Question

I’m using ECS Managed Auto Scaling with Target Tracking.

When the Target Tracking service performs a scale-out action, everything works as expected: the Auto Scaling Group (ASG) launches new instances to accommodate the increased demand.

However, when the Target Tracking service performs a scale-in action, the following issues occur:

• Tasks are stopped as expected.
• The Auto Scaling Group gets stuck in the "Updating capacity" state.
• The extra instances are not terminated, resulting in unused instances being left active.

This behavior seems inconsistent with the expected functionality, where the ASG should terminate unused instances during a scale-in event.

CloudWatch metrics show that the Target Tracking policy is triggering correctly.

Could you help me to figure out the following:

• What could cause the Auto Scaling Group to remain stuck in the "Updating capacity" state during a scale-in event?
• Are there additional configurations steps I should check to ensure the ASG properly terminates instances?

Answer 1

So, guys, the issue was caused by the absence of this tag in the Auto Scaling Group configuration:

resource "aws_autoscaling_group" "test" {
  # ... other configuration, including potentially other tags ...

  tag {
    key                 = "AmazonECSManaged"
    value               = true
    propagate_at_launch = true
  }
}