min_size = 1 results in errors when autoscaler attempts to remove the last instance

[lorengordon]

Using the autoscaler feature with min_size = 1, and seeing these errors in the lambda log when the autoscaler attempts to replace the last instance in the ASG.

{
    "time": "2024-03-14T14:24:05.448962118Z",
    "level": "WARN",
    "msg": "instance has no corresponding worker in Spacelift, removing from the ASG",
    "aws_request_id": "redacted",
    "asg_arn": "arn:aws:autoscaling:us-east-1:redacted:autoScalingGroup:redacted:autoScalingGroupName/redacted",
    "worker_pool_id": "redacted",
    "instance_id": "i-redacted",
    "launch_timestamp": 1710424905,
    "instance_age": 1340448951615
}

followed by:

{
    "errorMessage": "could not kill instance: could not detach instance from autoscaling group: operation error Auto Scaling: DetachInstances, https response error StatusCode: 400, RequestID: redacted, api error ValidationError: AutoScalingGroup redacted has min-size=1, max-size=50, and desired-size=1. To detach 1 instance, please update the AutoScalingGroup sizes appropriately.",
    "errorType": "wrapError"
}

The solution for this is for the min_size on the ASG to be set to 0, and for the desired_capacity to be set to 1.