Scheduling dependencies in DataWorks work by using the output of an upstream node as the input for a downstream node. This mechanism extends across business flows and workspaces, letting you ensure a downstream business flow starts only after its upstream counterpart succeeds. For more information, see Scheduling dependency configuration guide.
When to use cross-flow dependencies
If two business flows are tightly coupled and always run together, consider merging them into a single business flow. Cross-flow dependencies are appropriate when:
Two business flows run on different schedules.
Different teams own different business flows, but those flows share dependencies.
A business flow with multiple branches needs to complete fully before triggering a downstream flow.
Key concepts
Virtual node
A virtual node is a control node — a dry-run node that does not generate data. Use virtual nodes as the root node of a business flow to control scheduling order and execution timing. Virtual nodes also serve as aggregation points, collecting outputs from multiple branches so a single downstream dependency can be configured.
Configure dependencies across business flows
For business flows with multiple branches, use a virtual node as the aggregation point so the downstream flow has a single, reliable trigger.
The following example shows how to set up a cross-business flow dependency for two multi-branch business flows.
The setup uses four virtual nodes across the two business flows:
Business Flow 1 (upstream)
business_flow_1_start_virtual_node— start node for all branches in Business Flow 1business_flow_1_end_virtual_node— aggregation node; when it succeeds, Business Flow 1 is complete
Business Flow 2 (downstream)
business_flow_2_start_virtual_node— start node for all branches in Business Flow 2business_flow_2_end_virtual_node— aggregation node for Business Flow 2
Cross-flow dependency: Configure the output of business_flow_1_end_virtual_node as the input for business_flow_2_start_virtual_node. Business Flow 2 starts only after business_flow_1_end_virtual_node runs successfully.
DataWorks supports three ways to configure node dependencies: dragging with the mouse, manual configuration, and automatic parsing. In this example, use manual configuration — enter the output name of business_flow_1_end_virtual_node in the Upstream Dependencies configuration area of business_flow_2_start_virtual_node.
Related topics
Configure dependencies across workspaces
Cross-workspace dependencies work the same way as cross-flow dependencies: add the output of an upstream node in Workspace A as the input for a downstream node in Workspace B. The configuration steps are the same as for other scheduling dependencies. For details, see Configure same-cycle scheduling dependencies.
Limitations
| Limitation | Details |
|---|---|
| Same-region requirement | Both workspaces must be in the same region. |
| Older workspace compatibility | For some older workspaces, dependencies from a standard mode workspace to a basic mode workspace may not be supported. Submit a ticket to request a fix. |