Auto Scaling termination policies

Sure! Here’s a clean version of the Auto Scaling Group (ASG) termination policies explanation without any icons:

---

Default Termination Policy

If you don’t specify a termination policy, AWS uses the default behavior:

1. Select the Availability Zone (AZ) with the most instances.
2. From that AZ, select the instance with the oldest launch configuration or launch template.
3. If multiple instances qualify, pick the one closest to the next billing hour.
4. If still multiple, select one at random.

---

Custom Termination Policies

You can define a list of termination policies, and AWS will evaluate them in the order you provide. Here are the available policies:

Policy	Description
OldestInstance	Terminates the instance that has been running the longest.
NewestInstance	Terminates the instance that was most recently launched.
OldestLaunchConfiguration	Terminates instances launched using the oldest launch configuration.
OldestLaunchTemplate	Terminates instances launched using the oldest launch template.
ClosestToNextInstanceHour	Terminates the instance that is closest to the next billing hour.
AllocationStrategy	Applies when using mixed instance types with Spot and On-Demand. Optimizes based on price or capacity.
Default	Uses the built-in default logic.

---

Example (using AWS CLI)

aws autoscaling update-auto-scaling-group \
  --auto-scaling-group-name my-asg \
  --termination-policies "OldestInstance" "Default"

---

Common Use Cases

• For rolling updates, use OldestLaunchTemplate to phase out older versions.
• For cost efficiency, use ClosestToNextInstanceHour (more relevant for older billing models).
• For Spot instance strategies, use AllocationStrategy.
• To maintain newer instances and remove older ones, use OldestInstance.

---

Let me know if you’d like an example using Terraform or how to set it in the AWS Console.