This topic describes the billing rules for pay-as-you-go compute resources in MaxCompute.
How it works
MaxCompute pay-as-you-go compute resources scale on top of your subscription-based resources. You are charged based on the number of Compute Units (CUs) used and the duration of use. The system aggregates usage into CU-hours and bills you hourly at the level-1 quota level. Bill generation may be delayed, but your final bill will reflect the actual charges.
MaxCompute offers two types of pay-as-you-go compute resources: elastically reserved CUs and auto-elastic CUs. The key differences are outlined below:
Type | Scaling method | Billing and charges |
Elastically reserved CU | Scheduled scaling You scale elastically reserved CUs up or down on a fixed schedule that you define in Quota Plans and Time Plans. This allows you to automatically add a specified number of elastically reserved CUs to handle peak loads and remove them when the period ends. For more information about quota configuration, see Manage quotas. Important: You are charged for elastically reserved CUs while they are provisioned, even if they are not used. To avoid unnecessary costs, scale in these resources when they are no longer needed.
| Pay-as-you-go No upfront payment is required. You are charged based on the number of provisioned elastically reserved CUs × usage duration. Billing starts when the resources are scaled out and allocated to your subscription-based quota, and stops when they are scaled in. The system records usage data on the hour and whenever the number of CUs changes. Total usage for elastically reserved CUs is aggregated into CU-hours and billed hourly at the level-1 quota level.
|
Auto-elastic CU | Automatic scaling After you enable auto-scaling in a Quota Plan, the system automatically detects workload changes and scales resources up or down within a maximum limit that you define. You only need to set this limit to control costs, eliminating the need for manual resource provisioning and release. For more information about quota configuration, see Manage quotas. You are charged only for the resources you actually consume.
| Pay-as-you-go No upfront payment is required. The system monitors actual usage on a per-second basis and calculates charges based on the number of auto-elastic CUs consumed × actual usage duration. Total usage is aggregated into CU-hours and billed hourly at the level-1 quota level.
|
Note If you need to scale beyond 10,000 CUs, contact technical support for an assessment. For details, see Obtain online support.
Pricing
The following tables show the pay-as-you-go prices for elastically reserved CUs and auto-elastic CUs.
Type | Public cloud price | Remarks |
Elastically reserved CUs | USD 0.0488/CU/hour | If you use these resources for more than 16 hours a day, consider purchasing subscription-based reserved CUs, which is more cost-effective. |
Auto-elastic CUs | USD 0.0558/CU/hour | Best practices for auto-elastic CUs |