Serverless ApsaraDB RDS for SQL Server is now available for commercial use as of March 20, 2023. Designed for small and medium-sized enterprises and individual developers, it automatically scales CPU and memory in real time based on workload demand — you pay only for actual RCU consumption, billed hourly.
Effective date
March 20, 2023
What's available
| Item | Details |
|---|---|
| Regions | China (Hong Kong) and Singapore |
| Edition | RDS High-availability Edition |
| SQL Server versions | SQL Server 2016 Standard Edition, SQL Server 2017 Standard Edition, SQL Server 2019 Standard Edition |
-
Serverless RDS instances are available in the China (Hong Kong) and Singapore regions.
Create a Serverless instance in the ApsaraDB RDS console.
If you use a RAM user to create a Serverless RDS instance, attach the AliyunRDSFullAccess access policy to that RAM user. For more information, see Use RAM for resource authorization.
Service-Level Agreement (SLA) commitments do not apply to Serverless RDS instances or to the availability and liability of any relevant services. Alibaba Cloud is not liable for adverse impacts resulting from the use of Serverless RDS instances and relevant services.
Use cases
Serverless is best suited for workloads with intermittent or unpredictable demand where average compute utilization is low over time. The table below summarizes when to choose Serverless over provisioned instances.
| Serverless | Provisioned | |
|---|---|---|
| Workload pattern | Intermittent, unpredictable, with significant idle periods | Regular, predictable, with consistently high utilization |
| Compute scaling | Automatic | Manual |
| Billing granularity | Per RCU-hour (actual usage) | Fixed |
| Compute responsiveness | May vary after idle periods | Immediate |
Well suited for Serverless:
-
Dev/test environments — workloads with significant off-peak periods
-
Software as a Service (SaaS) applications — website hosting for small and medium-sized enterprises
-
Educational workloads — teaching environments and student experiments
-
Internet of Things (IoT) and edge computing — inconsistent, unpredictable workloads
-
Changing or unpredictable services — scenarios where service demand or configuration shifts frequently
-
Intermittent scheduled tasks — jobs that run periodically and stay idle between runs
-
Individual developers — variable personal projects
Less suited for Serverless:
-
Workloads with consistently high and predictable compute utilization
-
Applications that cannot tolerate any variability in compute responsiveness
Billing
Fees for a Serverless ApsaraDB RDS for SQL Server instance consist of two components:
Other fees are optional. For example, you are billed for basic backup storage costs for the serverless instance based on actual usage. For more information, see Billing.
Total fees = RCU fees + Storage costs + (Optional) Other fees
Other fees are optional. For example, basic backup storage is billed based on actual usage. For details, see Billing.
How RCUs are billed
Compute is measured in RDS Capacity Units (RCUs). Each hour, you are billed for the actual number of RCUs your instance uses, up to your configured maximum. The billable RCU range is 2 to 16 RCUs.
You are billed for the maximum storage capacity that you select at the time of purchase, even if your actual usage is lower. Storage cannot be scaled in after it has been scaled out. You are billed based on storage capacity after scale-out. To scale out storage, see Expand storage capacity.
Unit prices
|
Billable item |
Unit price |
|
RCU (RDS Capacity Unit) |
|
|
Storage costs |
ESSD PL1 disk: USD 0.0008 per GB-hour |
| Billable item | Unit price |
|---|---|
| RCU — China (Hong Kong) | USD 0.312 per RCU-hour |
| RCU — Singapore | USD 0.288 per RCU-hour |
| Storage (PL1 ESSD) | USD 0.0008 per GB-hour |
Billing example
The following example shows how to calculate daily fees for a Serverless instance in the Singapore region, configured with 2–8 RCUs and 200 GB of storage.
| Time period | RCUs in use | Duration | RCU-hours billed | RCU fees (USD 0.288/RCU-hour) |
|---|---|---|---|---|
| Peak hour | 8 RCUs | 1 hour | 8 RCU-hours | 0.288 × 8 × 1 = 2.304 |
| Off-peak | 2 RCUs | 23 hours | 46 RCU-hours | 0.288 × 2 × 23 = 13.248 |
| Total | — | 24 hours | 54 RCU-hours | USD 15.552 |
Storage costs: 0.0008 × 200 GB × 24 hours = USD 3.84
Total daily fees: USD 19.392
This example is for reference only. Actual fees are specified on your bill.
Billing example
For example, you purchase a serverless ApsaraDB RDS for SQL Server instance in the Singapore region. The instance has a compute specification of 2 to 8 RCUs and 200 GB of storage space. Over one day, the instance runs at 8 RCUs for 1 peak hour and at 2 RCUs for the remaining 23 hours. If the unit price is USD 0.288 per RCU-hour and the storage price is USD 0.0008 per GB-hour, the fees for that day are calculated as follows:
-
RCU fees: (0.288 × 1 × 8) + (0.288 × 23 × 2) = USD 15.552
-
Storage costs: 0.0008 × 24 × 200 = USD 3.84
Total fees = USD 15.552 + USD 3.84 = USD 19.392.
The preceding example is for reference only. Your actual fees are specified on your bill.