AWS supports a number of preferences that can control the behavior of an Instance Refresh operation: https://docs.aws.amazon.com/autoscaling/ec2/APIReference/API_RefreshPreferences.html.
Harness has already added support for some of these parameters, such as
MinHealthyPercentage
and
MaxHealthyPercentage
. We would like to be able to customize additional parameters, including:
  • StandbyInstances
    and
    ScaleInProtectedInstances
    : AWS's default behavior if these parameters are unspecified is to pause the instance refresh until an operator intervenes and terminates the standby instances or removes protection from the protected instances. The ability to customize these settings to other non-default options can allow the refresh to proceed without human intervention, avoiding Harness timing out the deploy step.
These additional options could either be configured in the ASG deploy step, or via an additional configuration file passed into the ASG Service definition.