Pay-as-you-go is a billing method that allows you to use resources first and pay for them afterward. Pay-as-you-go Elastic Compute Service (ECS) resources can be purchased and released on demand and provide cost savings of 30% to 80% compared with traditional hosts. This topic describes the billing and settlement rules for pay-as-you-go ECS resources.

Billing overview

Pay-as-you-go resources are billed on an hourly basis. If you have a quota agreement with Alibaba Cloud, fees are deducted only when the cumulative consumption of your account exceeds the quota. You must complete the payment at least once a month. Pay-as-you-go instances are suitable for services that experience spikes in traffic, such as temporary scaling, interim testing, and scientific computing.

You can change the configurations of your pay-as-you-go resources. For more information, see Change the instance type of a pay-as-you-go instance and Modify the bandwidth configurations of pay-as-you-go instances.

You can use one of the following methods to view the consumption details of your pay-as-you-go ECS resources:
  • For information about how fees are calculated, see the Billing examples section of this topic.
  • For information about how resource states affect the billing duration, see the Billing durations section of this topic.
    Note If you stop an instance but do not release its resources, you continue to be charged for the resources.
  • For information about settlement, see the Settlement cycles section of this topic.

Billable items

The combination of computing resources (vCPUs and memory), block storage devices, an image, and a network type uniquely determines the service form of an ECS instance. The pay-as-you-go billing method is applicable to the ECS resources described in the following table.
Resource Description
Computing resource (vCPUs and memory) When you create an instance, you must determine whether to use the pay-as-you-go billing method for the ECS resources involved.
Image The images created along with pay-as-you-go instances are billed on a pay-as-you-go basis.
Disk Disks created along with pay-as-you-go instances are billed on a pay-as-you-go basis.

After you create a pay-as-you-go instance, you can attach pay-as-you-go disks that were separately created to the instance. For more information, see Attach a data disk.

Public bandwidth (pay-by-bandwidth) If you select pay-by-bandwidth as the billing method for network usage when you create a pay-as-you-go instance, the amount of public bandwidth allocated to the instance is also billed on a pay-as-you-go basis. For more information, see Public bandwidth.
Snapshot

The billing cycle for a snapshot begins immediately after you create it. By default, the pay-as-you-go billing method is used for snapshots.

For more information, see Snapshots.

You can view the total price of the preceding resources in the lower part of the instance buy page in the ECS console.
  • Total indicates the total price of the following resources:
    • Computing resource (vCPUs and memory)
    • Cloud disks
    • Public bandwidth (pay-by-bandwidth)
      Note If you select pay-by-traffic as the billing method for network usage, the total price does not include the price of public bandwidth use. For more information, see Public bandwidth.
  • Marketplace Image Fees: If you selected a paid Alibaba Cloud Marketplace image, Marketplace Image Fees are displayed.

Settlement cycles

Pay-as-you-go ECS resources are billed on an hourly basis. They are paid together with other pay-as-you-go resources in your account. If you have a quota agreement with Alibaba Cloud, fees are deducted only when the cumulative consumption of your account exceeds the quota. If the cumulative monthly consumption by your account within the current month does not reach the quota, fees are deducted on the first day of the next month.
  • If your default payment method is bank card, the quota is USD 1,000.
  • If your default payment method is PayPal or Paytm (India), the quota varies based on your ECS usage.

The system attempts to deduct fees up to three times: on the due date (T), the 7th day after the due date (T+7), and the 14th day after the due date (T+14). If the fees for a pay-as-you-go instance cannot be deducted on the due date, payment becomes overdue for the instance and the system attempts to deduct the fees again on the 7th day after the due date (T+7) and on the 14th day after the due date (T+14). If all three attempts fail, the instance is stopped on the 15th day after the due date (T+15) and its billing also stops. For more information, see the "Pay-as-you-go resources" section in Overdue payments.

Billing durations

If a pay-as-you-go instance is automatically stopped due to an overdue payment, billing for some resources of the instance suspends. Billing for the resources resumes after you settle the overdue payment and reactivate the instance.

The following table describes the billing duration of each resource type for a pay-as-you-go instance on the premise that you have no overdue payments in your account.
Resource Billing duration
Computing resource (vCPUs and memory) The billing duration of computing resources (vCPUs and memory) is affected by the network type of the instance.
  • For an instance that resides in the classic network, billing begins when the instance is created and ends when the instance is released.
  • For an instance that resides in a virtual private cloud (VPC), the billing duration varies based on whether economical mode is enabled.
    • If economical mode is disabled, billing begins when the instance is created and ends when the instance is released.
    • If economical mode is enabled, billing begins when the instance is created and suspends when the instance is stopped by using the ECS console. Billing resumes when the instance is started again by using the ECS console and ends when the instance is released. For more information, see Economical mode.
      Important If you stop the instance by shutting down its operating system, economical mode cannot be not triggered.

You can purchase reserved instances to minimize costs. For more information, see Overview.

Image Billing begins when the instance is created and ends when the instance is released.
Cloud disk (system disk) Billing begins when the instance is created and ends when the system disk is released along with the instance.
Cloud disk (data disk) Billing begins when the data disk is created and ends when the data disk is released.
Public bandwidth (pay-by-bandwidth) Billing begins when public bandwidth (pay-by-bandwidth) is enabled and ends when public bandwidth is disabled or when the instance is released.

For information about how to disable public bandwidth, see Modify the bandwidth configurations of pay-as-you-go instances.

Snapshot Billing begins when the snapshot is created and ends when the snapshot is deleted.
Note If a pay-as-you-go instance incurs charges of less than USD 0.01 during its entire lifecycle, you are charged USD 0.01 for the instance.

To prevent unexpected charges incurred when a pay-as-you-go instance is not released in a timely manner, we recommend that you enable the automatic release feature for the instance. If automatic release is enabled, billing ends when the instance is released. The automatic release time is accurate to the second. For more information, see Enable automatic release.

Billing examples

You are charged for pay-as-you-go resources based on their billing cycles. The billing cycle of a pay-as-you-go resource begins immediately after the resource is created. You can calculate the total fee that you must pay for a period of time based on the configurations of your choice. The following table describes the billing cycle of each ECS resource and the formula used to calculate the fee for the resource.
Resource Billing cycle Formula Unit price
Computing resource (vCPUs and memory) Varies based on the number of the vCPUs of the instance type.
  • 1 vCPU: 10 minutes, with a minimum of 10 minutes
  • 2 vCPUs: 5 minutes, with a minimum of 5 minutes
  • 4 vCPUs: 2 minutes, with a minimum of 2 minutes
  • More than 4 vCPUs: 1 second
Unit price of an instance type × Billing duration For more information, see the Instance section on the Pricing tab of the Elastic Compute Service product page.
Note Local disks are tied to specific instance types. The prices of local disks are included in the prices of the instance types.
Image 1 second Unit price of an image × Billing duration You can view the price on the instance buy page in the ECS console or in Alibaba Cloud Marketplace.
Cloud disk (system disk) Varies based on the number of the vCPUs of the instance type.
  • 1 vCPU: 10 minutes, with a minimum of 10 minutes
  • 2 vCPUs: 5 minutes, with a minimum of 5 minutes
  • 4 vCPUs: 2 minutes, with a minimum of 2 minutes
  • More than 4 vCPUs: 1 second
Unit price of a cloud disk × Disk capacity × Billing duration For more information, see the Storage section on the Pricing tab of the Elastic Compute Service product page.
Note The prices for pay-as-you-go disks displayed on the tab are prices in USD per 100 GiB-hour. You can divide the prices by 100 to obtain the unit prices per GiB-hour.
Cloud disk (data disk) 1 second Unit price of a cloud disk × Disk capacity × Billing duration For more information, see the Storage section on the Pricing tab of the Elastic Compute Service product page.
Note The prices for pay-as-you-go disks displayed on the tab are prices in USD per 100 GiB-hour. You can divide the prices by 100 to obtain the unit prices per GiB-hour.
Public bandwidth (pay-by-bandwidth) 1 second Unit price of bandwidth × Bandwidth value × Billing duration

For more information, see Public bandwidth.

A tiered billing model is used for bandwidth. You can select a bandwidth value on the instance buy page to view the fee schedule.
Snapshot 1 hour, with a minimum of 1 hour Unit price of a snapshot × Snapshot capacity × Billing duration

For more information, see Snapshots.

For more information, see the Snapshot section on the Pricing tab of the Elastic Compute Service product page.
Note If the billing cycle is 1 second, you are charged for resources in 1-second increments. If an hourly price is displayed, you can divide the price by 3600 to obtain the unit price per second.
For example, assume that you created a pay-as-you-go instance in the China (Qingdao) region and used this instance from 11:00:00 to 12:00:00 on August 8, 2022. The following figure shows the process of calculating the instance price.
Note The prices in the following table are for reference only. Prices on the Pricing tab of the Elastic Compute Service product page prevail.

Change the billing method from pay-as-you-go to subscription

You can change the billing method of an ECS instance from subscription to pay-as-you-go. This allows you to recover a portion of the subscription costs and use the instance in a more flexible manner. For more information, see Change the billing method of an ECS instance from pay-as-you-go to subscription.

Overdue payments

When you have overdue payments within your account, pay-as-you-go resources cannot be used. Your pay-as-you-go instances may be stopped and the resources may be released. To prevent the consequences of overdue payments, such as instance stop or release from affecting your business, add funds to your account to complete the overdue payments at your earliest convenience. For information about changes in resource states when payments become overdue, see Overdue payments.

Billing FAQ