Skip to main content
Some Auto Scaling groups keep more instances running than their load needs, often because the minimum or desired capacity was set high. AWS recommends scaling these groups in to fewer instances. nOps shows these recommendations with AWS’s estimated monthly saving.

Why it matters

Every extra instance in a group is billed in full. Lowering the minimum or fixing a scaling policy removes capacity that is paid for but not used.

How nOps finds it

This recommendation comes from AWS (Compute Optimizer through Cost Optimization Hub), not from nOps. AWS looks at the group’s utilization over its lookback period and recommends a smaller number of instances. nOps imports the recommendation daily from AWS Cost Optimization Hub, which returns one recommendation per resource (the one with the highest saving), and shows AWS’s estimated monthly saving, which already takes your discounts into account.

Before you start

AWS Compute Optimizer and AWS Cost Optimization Hub enabled for the account (or the whole organization). Permission to update the Auto Scaling group (autoscaling:UpdateAutoScalingGroup).

How to implement

  1. Check why the group runs at its current size: minimum and desired capacity, scheduled actions and scaling policies.
  2. Lower the minimum and desired capacity (aws autoscaling update-auto-scaling-group --min-size N --desired-capacity N), or adjust the scaling policy so it scales in when load is low.
  3. Keep at least one instance per Availability Zone if the application needs to stay available during a zone outage.
  4. Watch latency and errors after the change.
  5. Update the capacity settings in Terraform or CloudFormation.

Risks and considerations

  • Fewer instances means less headroom for sudden load. Make sure scaling policies can add capacity fast enough.
  • To roll back, raise the minimum and desired capacity again.