Docs

Auto Scaling

Rolling Upgrade

Creating execution tasks

Scenarios

This topic describes how to create a rolling upgrade task. Based on your needs, you can configure the task to be executed in batches.

tip

Rolling upgrade supports all regions, but it is currently in Beta in the China (Shanghai) and China (Guangzhou) regions. To apply for a trial, contact your account manager. For more information, see Regions and availability zones.

Prerequisites

A rolling upgrade task can be executed only when the scaling group is in the Active status, and there are instances in the In service or Under protection status in the group.

If a rolling upgrade task is running for a scaling group, creating a new task is not allowed.

Usage notes

Once the task is created successfully, the rolling upgrade operation starts immediately. Evaluate the potential impact on your business before creating the task.

During the scaling group rolling upgrade, the scaling group process is automatically suspended and resumes automatically once the upgrade is complete.

During execution of a rolling upgrade task, instances change to Standby status, and those undergoing upgrade no longer receive business traffic. If the scaling group is associated with Load Balancer, the load balancing weight of the instances will be set to zero. Upon upgrade completion, instances automatically revert to pre-upgrade status and Load Balancer weight.

If the scaling group process is manually suspended before rolling upgrade, it remains suspended after the upgrade is complete. You are required to manually resume the suspended scaling group process. For detailed operations, refer to Pausing/resuming scaling group process.

If the upgrade fails or you manually cancel the task, you need to manually resume the scaling group process and remove instances from the "Standby" status.

Procedure

In the top menu bar, select the project and region of the target resource.

Click the name of the scaling group and select the "Rolling Upgrade" tab.

Click the "Create execution task" button and configure the parameters based on the table below.

Parameter	Description
Instance state	Select the status of instances to undergo rolling upgrade. Upon task creation, all instances in the scaling group in the specified status are automatically updated. Only instances in the In service or Under protection status support this feature.
Type	Update image: Used to update the image for instances in a scaling group. Updating an image has the following impacts on the instance configuration source of the scaling group: Scaling configuration: The image in the current scaling configuration associated with the scaling group is automatically updated. Images in other scaling configurations are not affected, and you can modify them manually based on your needs. Launch template: The image in the launch template is not updated automatically. You need to modify it manually to ensure the new image can be used for scaling out instances in the scaling group. Script execution: Used to perform O&M operations on instances in a scaling group, such as installing the Cloud Monitor plugin and changing instance passwords. tip Only Linux instances support this feature.
Image for updates/Script for execution	Select or enter the image/script to use for the update task. This image/script is selected by default for subsequent rollback tasks, and you can also choose other images/scripts. tip The selected image needs to support creating instances with the specifications configured in the scaling configuration source. Otherwise, the task execution fails. If the O&M assistant client was installed on an instance, it is not installed automatically during an image update. You need to install it manually after the update is complete.
Image/script for rollback	Select or enter the image/script to use for the rollback task. tip You can create a custom image based on an image currently used by an instance and use it as the rollback image. This allows instances to revert to the original configurations.
Execution batch	Number of batches for task execution. Each batch must contain at least one instance. For example, if six ECS instances in the scaling group are in In service status and the number of batches is set to 2, each batch contains three ECS instances, and the task runs in two batches.
Pause strategy	To confirm whether the instances that completed execution in each batch meet expectations, you can specify a pause strategy via this parameter. Do not suspend: The task is executed to completion in one run. Pause the first batch: After instances in the first batch complete the operation, the scaling group task is suspended, requiring you to manually resume the task. Operations for batches subsequent to the first are not suspended. Suspend each batch: The scaling group task is suspended after every batch is executed. You need to manually resume the task each time.

Click the "OK" button to complete the operation.

Subsequent operations

After a rolling upgrade task is executed, you can view the execution details, including the list of target instances and execution details.

Last updated: 2026.06.24 17:08:51