When your workload decreases, a scaling group initiates a scale-in event based on your configuration. During a scale-in, the scaling group reclaims instances based on the configured Instance Reclaim Mode. This topic describes how to configure the Instance Reclaim Mode.
Overview
When a scale-in event is triggered, the scaling group removes instances based on the configured Scaling Policy and Scale-In Policy. The removed instances are then reclaimed based on the specified Instance Reclaim Mode. The instance reclaim mode provides the following options.
-
This option is supported only if the scaling group type is ECS and the network type is VPC.
-
If the scaling group type is ECI, you cannot configure the Instance Reclaim Mode. For ECI scaling groups, the default Instance Reclaim Mode is Release.
-
Release
If you select this mode, during a scale-in, the system first stops the instance and then releases it. If the stop operation times out, the instance is forcibly released. No resources are retained. During a scale-out, Auto Scaling creates new instances and adds them to the scaling group.
-
Force Release
If you select this mode, during a scale-in, the system forcibly releases instances that are in the Running (
Running) state. Forcibly releasing an instance is equivalent to unplugging its power source, which erases all data from its memory and ephemeral storage. This data cannot be recovered. Use this mode with caution. -
Economical Mode
In this mode, the system stops the removed ECS instances, which in turn triggers the Economical Mode feature. Some resources are retained and continue to be billed. During a scale-out, Auto Scaling prioritizes reusing these stopped instances before creating new ones. This mode can improve scaling efficiency. For more information, see Use Economical Mode to improve scaling efficiency.
-
Force Economical Mode
If you select this mode, during a scale-in, the system forcibly shuts down instances that are in the Running (
Running) state. A forced shutdown is equivalent to unplugging the instance's power source, which erases all data from its memory and ephemeral storage. This data cannot be recovered.
-
To prevent data loss when an instance is reclaimed, do not store application data or logs on the instance.
-
Stopped instances may be released in the following scenarios:
-
If you manually trigger a scale-in by reducing the Maximum Number of Instances for the scaling group, Auto Scaling prioritizes the release of stopped ECS instances.
-
If a stopped instance cannot be added to the scaling group due to insufficient inventory or an overdue payment, Auto Scaling releases the instance.
-
-
The behavior of Economical Mode is affected by the Economical Mode feature for pay-as-you-go instances. For more information, see Economical Mode.
Configure the instance reclaim mode
Existing group
To modify the Instance Reclaim Mode for an existing scaling group, follow these steps.
-
Log on to the Auto Scaling console.
In the left-side navigation pane, click Scaling Groups.
In the top navigation bar, select the region where Auto Scaling is activated.
-
On the Scaling Groups page, find the scaling group for which you want to modify the Instance Reclaim Mode, and click Details in the Actions column.
-
Find the Instance Reclaim Mode parameter and select a mode.
-
Click OK.
New group
When you create a scaling group, configure the Instance Reclaim Mode as shown in the following figure.
