Hybrid billing lets you add elastic compute capacity on top of a fixed subscription quota. Use it when your workloads have a stable, predictable baseline but experience temporary usage spikes — for example, batch jobs that run during business hours with occasional end-of-month surges.
The subscription quota cost is subject to the subscription payment page or the final bill. For the elastic computing resource cost, see your bill. For more information, see Query bills.
For example, consider a workspace in the China (Hong Kong) region with a maximum pay-as-you-go quota of 100 Compute Units (CUs) at USD 0.074048 per CU per hour, and a subscription quota of 50 CUs at USD 35.54 per CU per month. The workspace runs at 100 CUs for one peak hour each day and does not exceed 50 CUs at other times. The estimated monthly cost is calculated as follows:
Subscription quota fee:
50 CU * 35.54 USD/CU/month * 1 month = 1777 USD.Elastic resource cost: Pay-as-you-go charges apply only when the workspace's actual usage exceeds the subscription quota during that period.
Daily pay-as-you-go usage:
100 CU * 1 hour - 50 CU * 1 hour = 50 CU*hours.Pay-as-you-go fee:
50 CU*hour * 0.074048 USD/CU/hour * 30 = 111.072 USD.
The estimated cost for this workspace is USD 1,777 + USD 111.072 = USD 1,888.072.
Limits
Hybrid billing can be enabled only on subscription workspaces in the Running state.
A hybrid billing workspace cannot be converted back to a pure subscription or pay-as-you-go workspace.
The hybrid billing queue feature is an allowlist feature. Submit a ticket to request access before you enable it.
Billing rules
Elastic compute resources are billed only for actual usage, charged per 1-hour cycle.
Cost formula:
| Cost component | Formula |
|---|---|
| Subscription quota cost | Number of CUs × Unit price per CU × Subscription duration |
| Elastic resource usage cost | Actual elastic CUs consumed × Unit price per CU × Hours used |
Total: Subscription quota cost + Elastic resource usage cost
The unit price and calculation method for elastic resources follow the pay-as-you-go rules for Serverless Spark. To verify actual charges, see Query bills.
Example (China (Hong Kong) region)
Assume a subscription quota of 100 Compute Units (CU) at USD 35.54/CU/month for 1 month, an elastic unit price of USD 0.074048/CU/hour, and actual elastic usage of 50 CU for 1 hour/day during peak hours over 30 days:
Subscription quota cost:
100 CU × USD 35.54/CU/month × 1 month = USD 3,554Elastic resource cost:
50 CU × USD 0.074048/CU/hour × 1 hour × 30 days = USD 111.072Estimated total: USD 3,665.072
For subscription pricing in other regions, see Subscription.
Enable hybrid billing
Prerequisites
Before you begin, ensure that you have:
A subscription workspace in the Running state
Allowlist access for the hybrid billing queue feature (submit a ticket if not already granted)
Steps
Log on to the E-MapReduce console. In the left navigation pane, choose EMR Serverless > Spark.
On the Spark page, find the target subscription workspace. In the Actions column, click
> Enable hybrid billing.In the dialog box, set the Maximum Pay-as-you-go Quota and click OK. After enabling, note the following:
The Billing Method for the workspace changes from Subscription to Mixed Payment.
Elastic compute resources are allocated to the pay_as_you_go queue by default. To change this allocation, go to the Queues tab on the Resource page. For details, see Manage resource queues.
Submit Spark jobs to a queue that has elastic compute resources assigned.
View elastic resource usage
Log on to the E-MapReduce console. In the left navigation pane, choose EMR Serverless > Spark.
On the Spark page, click the target workspace.
In the left navigation pane, click Resources. On the Queues tab, view elastic compute resource allocation and usage.
To review billing details, see Query bills.