E-MapReduce (EMR) supports subscription and pay-as-you-go billing. Subscription is a prepaid model: you pay upfront for a fixed term and reserve capacity for the duration. For workloads with predictable, long-running resource needs, subscription is more cost-effective than pay-as-you-go in most business scenarios.
When to use subscription
Subscription suits workloads that meet all of the following conditions:
The usage period is predictable.
Resource demand is stable, without frequent spikes or drops.
Resources are needed for an extended period.
For short-lived or variable workloads, use pay-as-you-go instead.
Billing details
| Item | Details |
|---|---|
| Billing rule | Prepaid. Pay before use, with resource reservation for the committed term. |
| Pricing | View the full price breakdown in the order generated when you purchase an EMR cluster in the EMR console. |
| Billing cycle | Starts at the exact second the subscription resource is purchased or renewed (UTC+8). Ends at 00:00:00 on the day after the expiration date. The unit is calendar months or years. |
| Refunding | Refunds are available for clusters that have not yet expired. For details, see Refund policy. |
| Overdue payments | For details, see Overdue payments. |
Billing cycle example
You purchased a subscription EMR cluster at 10:23:30 on January 1, 2023, with a three-month term. The cluster expires at 00:00:00 on April 2, 2023.
The billing cycle unit is calendar months or years, not fixed 30-day periods. This means the end date depends on the actual calendar.
Pay-as-you-go resources in a subscription cluster
Some resources within a subscription cluster are billed on a pay-as-you-go basis:
Pay-as-you-go resources in the cluster: If the cluster includes pay-as-you-go resources, bills are generated for the associated ECS instances and EMR service usage. For details, see Pay-as-you-go.
Auto scaling nodes: If node groups in the cluster have auto scaling rules configured, ECS instances added by auto scaling and the associated EMR service usage are billed on a pay-as-you-go basis. For more information about auto scaling rules, see Configure custom auto scaling rules.
Impact after instance expiration
When instances in a subscription EMR cluster expire, the following operations are restricted:
Service operations: adding, restarting, stopping, and configuring parameters
User management operations
Auto scaling operations
Node operations: scaling and configuration upgrade
Viewing application performance monitoring data
Bootstrap action and script operations