DataWorks no longer recommends exclusive resource groups for scheduling. Use Serverless resource groups instead. Serverless resource groups consolidate the capabilities of exclusive resource groups for scheduling, exclusive resource groups for Data Integration, exclusive resource groups for DataService Studio, and shared resource groups for scheduling into a single resource group type. You can use a single Serverless resource group to synchronize data, schedule and run tasks, and call and manage API services.
Exclusive resource groups for scheduling provide dedicated Elastic Compute Service (ECS) instances for running DataWorks tasks. Unlike shared resource groups, these groups give your tasks isolated computing capacity — preventing resource contention from affecting task timeliness.
Use cases
Exclusive resource groups for scheduling are suited for the following scenarios:
Time-sensitive tasks: Your tasks have strict scheduling deadlines and cannot tolerate delays caused by resource contention in shared resource groups.
Flexible capacity: You need to scale the resource group up or down based on workload changes.
Network access: Your tasks need to reach the Internet, Virtual Private Clouds (VPCs), or Internet Data Centers (IDCs).
Access control: You need IP address whitelists to restrict which resources can connect to your data sources.
Task type requirements: Some DataWorks task types — including E-MapReduce (EMR), CDH, and ADB tasks — support only exclusive resource groups.
Limitations
Purchase and release constraints (affect your buying decision):
The billing method is subscription. You cannot delete or release a resource group before it expires. After expiry, the service is suspended and the resource group is released after a specified period.
Exclusive resource groups for scheduling cannot be used across regions. For example, a resource group in the China (Shanghai) region can only be used by workspaces in China (Shanghai).
Operational constraint (affects day-to-day use):
You cannot run the
yumcommand inside an exclusive resource group for scheduling.
Specifications and concurrency
| Specifications | Maximum concurrent instances |
|---|---|
| 4 vCPU 8 GiB | 16 |
| 8 vCPU 16 GiB | 32 |
| 12 vCPU 24 GiB | 48 |
| 16 vCPU 32 GiB | 64 |
| 24 vCPU 48 GiB | 96 |
Billing
Task runtime costs on an exclusive resource group are included in the resource group subscription fee and are not charged separately to shared resource groups.
The billing method is subscription. Pricing varies by region and specification:
| Region | 4 vCPU 8 GiB | 8 vCPU 16 GiB | 12 vCPU 24 GiB | 16 vCPU 32 GiB | 24 vCPU 48 GiB | 32 vCPU 64 GiB |
|---|---|---|---|---|---|---|
| China (Hangzhou) | $76.23 | $152.45 | $228.68 | $304.91 | $457.36 | $609.82 |
| China (Shanghai) | $76.23 | $152.45 | $228.68 | $304.91 | $457.36 | $609.82 |
| China (Shenzhen) | $76.23 | $152.45 | $228.68 | $304.91 | $457.36 | $609.82 |
| China (Beijing) | $76.23 | $152.45 | $228.68 | $304.91 | $457.36 | $609.82 |
| China (Zhangjiakou) | $57.27 | $114.34 | $171.61 | $228.68 | $343.02 | $457.36 |
| China (Chengdu) | $62.49 | $124.79 | $187.09 | $249.38 | $374.17 | $498.77 |
| US (Silicon Valley) | $149.26 | $298.53 | $447.77 | $597.06 | $895.55 | $1,195.00 |
| US (Virginia) | $110.63 | $221.26 | $331.89 | $442.53 | $663.78 | $885.06 |
| China (Hong Kong) | $138.73 | $276.58 | $416.18 | $553.16 | $832.35 | $1,106.32 |
| Singapore | $138.73 | $276.58 | $416.18 | $553.16 | $832.35 | $1,106.32 |
| Germany (Frankfurt) | $138.73 | $276.58 | $408.96 | $553.16 | $817.91 | $1,106.32 |
| Malaysia (Kuala Lumpur) | $130.01 | $259.06 | $390.05 | $517.92 | $780.07 | $1,035.84 |
| Indonesia (Jakarta) | $138.73 | $276.58 | $416.18 | $553.16 | $832.35 | $1,106.32 |
| Japan (Tokyo) | $158.92 | $316.97 | $474.14 | $633.94 | $948.27 | $1,268.75 |
| UK (London) | $104.67 | $209.33 | $314.00 | $418.67 | $628.00 | $837.14 |
| UAE (Dubai) | $166.38 | $331.80 | $498.38 | $663.80 | $996.76 | $1,327.59 |
| SAU (Riyadh - Partner Region) | $166.48 | $331.90 | $499.42 | $663.79 | $998.82 | $1,327.58 |
Prices are in USD per month.
Scale-out and scale-in
When you purchase a resource group, specify the machine specifications and the number of nodes. After purchase, you can scale out (add nodes) or scale in (remove nodes) at any time. For the procedure and considerations, see Scale out or scale in a resource group.
Change specifications
If the current node specifications no longer fit your workload, you can upgrade or downgrade. Changing specifications applies to all nodes in the resource group simultaneously. For the procedure and considerations, see Change specifications.
Renewal and release
Renew the resource group before it expires, or let it expire and be released automatically. For renewal options, see Expiration and renewal.
After a resource group is released, all tasks that rely on it stop running.
Network connectivity
If your tasks do not interact with any data source, no network configuration is needed — the resource group is ready to use immediately after purchase.
After you purchase an exclusive resource group for scheduling, you must attach it to a VPC. Then, configure connectivity based on where the data source resides.
| Data source location | Connectivity solution |
|---|---|
| Internet | Connect the data source directly to the VPC attached to the resource group. |
| VPC in the same region | Attach the resource group to the same VPC as the data source. The system automatically adds the required route. |
| VPC in a different region | Connect the two VPCs using Express Connect or VPN, then add a route pointing to the destination database IP address. |
| On-premises data center | Connect the data center to the resource group VPC using Express Connect or VPN, then add a route pointing to the destination database IP address. |
| Classic network | Exclusive resource groups reside in a VPC and cannot connect to classic network resources. Migrate the data source to a VPC first. |
Whitelist configuration
If the data source restricts access by IP address whitelist, add the endpoint of the exclusive resource group to the data source's whitelist. For instructions, see Add a whitelist.
FAQ
For answers to common questions, see Exclusive resource groups FAQ.