Every API call in DataService Studio runs on a resource group -- a set of servers that provide CPU, memory, and bandwidth for query execution. The resource group you choose directly affects API reliability, latency, and throughput.
How resource groups work
When a client calls an API, the request follows this path:
API Gateway receives the request.
The request is routed to a DataService Studio server in the resource group.
The server queries the destination data source and returns the result.
The resource group handles both network routing and query execution. Two factors determine whether API calls succeed:
Connectivity -- The resource group must reach the data source over the network. If the network path is broken, API calls fail.
Capacity -- API calls consume CPU, memory, and network resources. Insufficient resources lead to call exceptions, throttling, or delayed query results.
Resource group types
DataService Studio supports two categories of resource groups: exclusive and public.
Exclusive resource group (recommended)
An exclusive resource group is purchased for your dedicated use. No other tenants share its capacity. Two options are available:
Serverless resource group (recommended) -- For details, see Use Serverless resource groups.
Exclusive resource group for DataService Studio -- For details, see Exclusive resource groups for DataService Studio.
NoteExclusive resource groups for DataService Studio are available only in the China (Shanghai) region.
Public resource group (for testing only)
A public resource group is provided by DataWorks and shared across all tenants. During peak hours, tenants compete for resources, which can degrade performance. For details, see Use a public resource group.
Compare resource group types
The following table compares exclusive and public resource groups.
Dimension | Exclusive resource group | Public resource group |
Ownership | Maintained by DataWorks, exclusively used by each tenant | Maintained by DataWorks, shared among all tenants |
Network connectivity | Internet, Alibaba Cloud virtual private clouds (VPCs) in instance and connection string modes, and data centers | Internet, Alibaba Cloud VPCs in instance mode, and classic network. Note Cannot access classic network data sources in the China (Shenzhen) Finance Cloud. |
Billing | Subscription, charged by resource group specifications | Pay-as-you-go tiers, charged by number of calls and call duration |
Supported data sources | ClickHouse, Hologres, RDS, MySQL, PostgreSQL, SQL Server, Oracle, Tablestore, AnalyticDB for MySQL 2.0, AnalyticDB for MySQL 3.0, AnalyticDB for PostgreSQL, MongoDB, DRDS, StarRocks, and Doris. More data sources will be supported in the future. | Hologres, RDS, MySQL, PostgreSQL, SQL Server, Oracle, Tablestore, AnalyticDB for MySQL 2.0, AnalyticDB for MySQL 3.0, AnalyticDB for PostgreSQL, MongoDB, and DRDS |
Maximum queries per second (QPS) | Varies by specification. Serverless resource groups: minimum 125 QPS. Exclusive resource groups for DataService Studio: minimum 500 QPS. One exclusive resource group can be attached to multiple workspaces and APIs. | 200 QPS per tenant per region. To increase the QPS threshold, switch to an exclusive resource group. |
Throttling behavior | If API calls exceed the QPS threshold, throttling is triggered and calls fail. | If API calls exceed 200 QPS, throttling is triggered and calls fail. |
Timeout limit | 30 seconds | 10 seconds |
Reliability | High | Low |
Security | High | High |
Best for | Highly concurrent, frequent online API calls with complex queries and large result sets | Low-concurrency or low-frequency API calls |
Rating | Recommended | Testing only |
-
You can no longer purchase resource groups of the api.s1.small, api.s1.medium, or api.s1.large specifications. You can only upgrade, downgrade, or release existing resource groups of these types of specifications.
-
Note1: The maximum queries per second (QPS) for exclusive resource groups is calculated based on actual business scenarios. You can estimate the QPS threshold by using the following information:
-
Whether to generate an API in script mode.
-
Whether the pagination feature is enabled for an API call so that the returned results are displayed on multiple pages.
-
The average runtime of SQL statements configured for an API call is 100 milliseconds in a data source.
-
The average size of data returned by a single API call is 3,000 bytes.
If your business scenario is different from the preceding scenario, join the DataWorks DingTalk group to obtain the appropriate specifications that suit your business scenarios.
-
Use an exclusive resource group for production workloads. Public resource groups are suitable only for testing.
Configure network connectivity
Before making API calls, make sure the resource group can reach the data source over the network. The configuration steps differ by resource group type.
Serverless resource group
Make sure the Serverless resource group can connect to the target data source. For configuration details, see Network connectivity configuration.
Exclusive resource group for DataService Studio
Select a network connectivity solution based on the network environment of the data source and configure a whitelist. For details, see Obtain IP addresses for a whitelist: Exclusive resource groups for DataService Studio.
Public resource group
Add the CIDR blocks of the corresponding region to the whitelist of the data source. For details, see Obtain IP addresses for a whitelist: Public resource groups.
Configure a DataService Studio quota for a Serverless resource group
To use a Serverless resource group with DataService Studio, allocate a compute unit (CU) quota for DataService Studio on the resource group before setting it as an Exclusive Resource Group for DataService Studio.
Without a configured quota, the following issues occur:
The connectivity test between the data source and the DataService Studio resource group fails.
The Serverless resource group does not appear as an option when configuring an exclusive resource group for DataService Studio.
Go to the Resource Groups page.
Log on to the DataWorks console. In the top navigation bar, select the desired region. In the left-side navigation pane, click Resource Group to go to the Resource Groups page.
In the Actions column of the Serverless resource group, click the
icon and select Manage Quota. To allocate resources to DataService Studio, set Occupied CUs for a pay-as-you-go resource group or Minimum CUs for a subscription resource group.Pay-as-you-go resource group
Subscription resource group
After you allocate the CUs, click OK to save the configuration.
Go to the DataService Studio page and configure the exclusive resource group again.
