Docs

Auto Scaling

Scaling rule

Creating scaling rules specific to ESIs

Scenarios

Elastic Scheduled Instances (ESI) are pay-as-you-go instances that support advance reservation. They are designed for elastic usage scenarios to ensure stable resource supply. After you place an ESI order, the platform automatically creates the specified number of instances within the resource delivery period. For more information, refer to Elastic Scheduled Instances (ESI).

tip

The feature of Elastic Scheduled Instances (ESI) in AS is in Beta. If you want to apply for a trial, please contact your account manager.

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 "Scaling Rules" tab.

Click "Create Scaling Rules" and refer to the table below to configure the parameters

tip

To create scaling rules for ESI orders only, set the rule type to "Scheduled task" or "Periodic task".

The following only describes the required parameters for ESI order-related scaling rules. All other parameter configurations are the same as those for standard scaling rules. For more details, refer to Creating scheduled scaling tasks and Creating periodic scaling tasks.

Parameter	Description	Sample value
Scope of action	Current scaling group: Indicates creating a regular scaling rule. For ESI orders only: Indicates creating an ESI order-related scaling rule. The system automatically creates one or multiple pay-as-you-go instances within the resource delivery period. Scaling rules dedicated to ESIs can be created only when the availability zone and instance type of the selected scaling configuration source support the reservation purchase feature. What instance types support scheduled purchases?	For ESI orders only
Reservation type	Indicates the type of ESI instances, including general-purpose and segmented types. This parameter defaults to the ESI type for which the allowlist permission is enabled in the availability zone of the selected scaling configuration source. Manual modification is not supported.	General-purpose
Number of reserved instances	Specifies the quantity of instances to be created in one task. Instances will be automatically delivered within this delivery time period.	3
(First) delivery time period	Specifies the time period for the first instance delivery (scheduled tasks are delivered only once), accurate to the minute. Each 10-minute segment is a separate time period. You can select a time period within 30 days. After selection, your instances will be delivered within this time period. For example, if the current time is 2024-02-27 17:42, the earliest first delivery time period supported is 2024-02-27 21:50 to 22:00. For periodic tasks, the time configured here must be earlier than the recurrence end time. Otherwise, the scaling task fails to execute. Instances delivered at a scheduled time follow a minimum 1-hour consumption policy.	2024-02-27 21:50 - 22:00
(First) submission time of ESI order	The adjustment of submitting the ESI order at the specified time is triggered for the scaling group. Then, the ESIs will be delivered during the specified delivery period. Selection is limited to 4 to 8 hours before the first delivery time slot. For example, if the time slot of the first scheduled delivery is 2024-02-27 21:50 to 22:00, then the submission time of the ESI order needs to fall within the range of 2024-02-27 13:50 to 17:50.	2024-02-27 17:42
Best-effort delivery	Indicates that when the inventory of Elastic Scheduled Instances (ESI) cannot fully meet the reservation requirements, it reserves as many instances as possible. Selecting this option means that the minimum quantity of instances is 1. When 1 ≤ inventory ≤ reservation quantity, reservation orders will be created based on the inventory. When inventory < 1, the creation of a reservation order will fail due to insufficient inventory.	Unchecked

Click the "Confirm" button to complete the operation.

Usage limit for related features

After the allowlist permission for Elastic Scheduled Instances (ESI) is enabled, logic changes of related features of scaling groups are as follows. Please read carefully:

Feature	Operation	Logic change
Create scaling groups	Create scaling groups	When you select a version of the instance launch template with the billing method "ESI", the billing method is automatically changed to "Pay-as-you-go".
Create scaling groups		Disabling scaling groups, Deleting scaling groups	You are not allowed to disable or delete scaling groups with ESI orders submitted but without instances delivered. To disable or delete a scaling group, cancel the ESI order first, and then proceed with the operation.
Scaling configuration source	Associating instance launch templates	When you associate a version of the instance launch template with the billing method "ESI", the billing method is automatically changed to "Pay-as-you-go". Scaling rules dedicated to ESIs can be created only when the availability zone and instance type of the selected scaling configuration source support the reservation purchase feature. What instance types support scheduled purchases?
Instance management	Modifying expected instance count or capacity	After instances in an ESI order are created, reclaimed, or deleted, the expected number of instances is automatically modified.
Instance management		Modifying instance count or capacity of scaling group	When you change the expected number of instances, the sum of the modified expected number of instances and the number of instances specified in the ESI order must not exceed the maximum number of instances in the scaling group. When you adjust the maximum number of instances for a scaling group, the adjusted maximum number must not be less than the number of instances specified in the ESI order.

Subsequent operations

After the scaling rule is triggered, you can view the scaling result details via the scaling log, including the conditions that triggered scaling, adjustment type, start/stop time, and execution result.