Docs

Elastic Compute Service

Elastic Scheduled Instance

Elastic Scheduled Instances (ESI) Overview

Elastic Scheduled Instances (ESI) are pay-as-you-go instances that support advance reservation, are designed for elastic scenarios, and ensure stable resource supply. After you make a reservation purchase, a Pay-as-you-go instance can be automatically created for you within the resource delivery time window.

tip

Currently, only accounts that have completed enterprise real-name authentication can use this feature. How do I upgrade a personal authentication to an enterprise authentication?

Elastic Scheduled Instances are currently only available for sale in the China (Beijing), China (Shanghai), and China (Guangzhou) regions. Moreover, the instance types offered in different regions and availability zones vary slightly. Please refer to the actual display in the console for details.

Billing rule

Elastic Scheduled Instances (ESI) are subject to a minimum one-hour charge. Regardless of whether the instance is released within the first hour after creation, a bill will be issued and a fee for one hour of usage will be charged. If usage exceeds one hour, charges will be based on actual usage. For detailed billing rules for Elastic Scheduled Instances, refer to Elastic Scheduled Instances (ESI).

Process description

Create an ESI order

When making a reservation, you need to complete the instance configuration (including compute specifications, image, storage, network, number of instances, and so on), and specify the resource delivery time window. After a reservation is successful, the system will automatically generate an ESI order. Learn more.

tip

After a reservation order is successfully created, it is not recommended to modify the relevant configuration of a pending delivery reservation order. Otherwise, instance delivery may fail. For details, refer to Precautions.

You can cancel the appointment within 30 minutes after the appointment order is created.

Automatic delivery of Elastic Scheduled Instances (ESI) by the system

After the resource delivery time window specified in the reservation order is reached, the system automatically creates a Pay-as-you-go instance based on the reservation order information. At this time, the instance status is Creating.

Instance delivery results are as follows:

Delivered: All instances in the reservation order have been created, and the instance status has changed to Running.

Partial delivery: There is a delivery risk, and some instances failed to be created. You can view details through event notifications sent by the platform via channels such as internal message, SMS, email, and so on.

Delivery failed: There is a delivery risk; all instances failed to be created. You can view details through event notifications sent by the platform via channels such as internal message, SMS, or email.

Release an instance

After Elastic Scheduled Instances (ESI) are successfully delivered, they will not be released automatically (except when resources are released due to an overdue payment). You can manually delete the instance after your business is complete.

Limits

Limitations	Description
Instance type	Instance families that support scheduled purchase: General Compute-optimized Memory-optimized HFG2 general-purpose instance with high processor base frequency Local-SSD i3s and i2 Some heterogeneous compute-optimized instance families
vCPU quota	A single account can use up to 300 vCPUs in a single availability zone. If you need to increase your vCPU quota, please go to the Quota Center to apply. Learn more.
Number of instances	You can purchase up to 500 instances at a time.
Deduction on the bill	Using Reserved Instances (RI) to deduct the compute resource (vCPU and memory) portion of the Elastic Scheduled Instances (ESI) bill is not supported; however, Reserved Storage Capacity can be used to deduct the storage portion of the bill.
Modify instance type	Not supported.
Convert billing methods	Not supported.
Pay-as-you-go Economical Mode	Not supported.
Deployment Set	Not supported.

Precautions

Due to the potential for delayed delivery with Elastic Scheduled Instances (ESI), it is not recommended to perform any actions that modify resources associated with the reservation order from the time the order is created until resource delivery is complete, as this may result in delivery failure, including but not limited to the following actions:

Time periods	Resource	Risk actions
From the start of reservation order creation to the completion of resource delivery	Network Interface	When creating a reservation order, a network interface is pre-created (at this point, the network interface is not yet mounted to the instance). The following actions are not allowed on this network interface; otherwise, resource delivery will fail. Mounting to other instances or deleting the network interface is not allowed. Changing the security group associated with the network interface, or associating or disassociating an EIP, is not permitted. You are not allowed to delete resources attached to the network interface, such as VPCs, subnets, security groups, and other resources. If an IPv6 address has already been assigned to the network interface, you are not allowed to enable IPv6 Internet bandwidth for it.
		Bandwidth Package	Pay-as-you-go Shared Bandwidth Packages: cannot be deleted. Subscription Shared Bandwidth Packages must not be in Expired, Suspended, or Unsubscribed status.
		Safety package	If an enhanced protection type EIP that supports DDoS Native Protection is associated, the security protection package must be in normal status. For more information about security protection, refer to the EIP part of Network configuration.
		Images	For ESI orders created using a Custom Image, deleting the Custom Image is not allowed. For ESI orders created using a Shared Image, canceling the Shared Image is not allowed.
		Key Pair	You are not allowed to delete key pairs associated with the instance before delivery.
		Project	Deleting projects associated with the reservation order is not allowed before instance delivery.
		Other risks: The account is overdue. Instance delivery failed. You can refer to bill management to view overdue details and top up promptly. Other actions that may modify the resources relied upon by Elastic Scheduled Instances (ESI), as well as instance delivery failures caused by storage or network issues, such as insufficient Volume inventory, insufficient EIP inventory, and so on.

Application scenarios

A foreseeable peak in one-off resource demand

In business scenarios where a large volume of access is expected within a short period, it is necessary to reserve resources for critical business during specific time windows and maintain high-performance access. For example: major e-commerce promotions, video live streaming, peak ticketing periods during holidays, and so on.

Regular periodic business peaks

For predictable computing workloads that are scheduled to occur at fixed daily, weekly, or monthly intervals, where application load peaks are concentrated within these periods, Elastic Scheduled Instances (ESI) can be used to ensure stable resource supply and an optimal user experience during peak times. For example, daily business peak periods for online gaming, online education, big data analytics, and rendering services, as well as monthly reconciliation peak periods for financial services and so on.

Elastic requirements for offline businesses that can tolerate some delay

Business peak periods are irregular and difficult to predict. However, for offline workloads that can tolerate some waiting time, you can schedule Elastic Scheduled Instances (ESI) once a business peak begins and run your workloads after the instances are delivered. This approach provides lower costs and greater resource certainty compared to Pay-as-you-go billing. For example: irregular business peaks in scientific computing, offline rendering, and so on.

Lifecycle

ESI order status

After the reservation is successful, the system generates an ESI order and, within the resource delivery time window, delivers instance resources according to the order.

Status name	Status attribute	Status description
Creating	Intermediate status	The status after you create an ESI order and before it enters Pending delivery.
Pending	Stable status	The normal status of an ESI order after it is created and before resource delivery.
Delivering	Intermediate status	After the resource delivery time period is reached, resource delivery begins.
Delivered	Stable status	All resources have been delivered.
Failed	Stable status	All resources failed to be delivered.
Partially failed	Stable status	Not all resources have been delivered.
Invalid	Stable status	The ESI order record will be permanently deleted 15 days after all resources have been released and the ESI order has become invalid.
Canceled	Stable status	Cancel the reservation.

Elastic Scheduled Instance statuses

After the successful delivery of Elastic Scheduled Instances (ESI), their lifecycle is the same as that of regular ECS instances. For details, refer to instance lifecycle.

Usage guide

Actions	Document
Create an ESI order	Create Elastic Scheduled Instances
View Elastic Scheduled event	View ESI order event
Purchase reservation orders with the same configuration	Purchase reservation orders with the same configuration
Purchase Elastic Scheduled Instances (ESI) with the same configuration	Purchase instances with the same configuration
Cancel appointment	Cancel appointment
Stop Elastic Scheduled Instances (ESI)	Stop instances
Delete Elastic Scheduled Instances (ESI)	Delete Elastic Scheduled Instances (ESI)

FAQs

Elastic Scheduled Instances (ESI) FAQ.

Last updated: 2026.06.18 16:22:23