After configuring scheduling dependencies for an auto triggered node, preview the dependency graph (DAG) to confirm the configuration is correct before deploying. The preview also shows the estimated execution time of instances generated on each scheduling day, so you can catch misconfigurations early and avoid delayed task runs.
Key concepts
Two distinct types of dependency relationships exist in a scheduling DAG:
| Type | Description | Visual indicator |
|---|---|---|
| Same-cycle dependency | A node depends on upstream nodes within the same scheduling cycle. | Solid lines |
| Cross-cycle dependency | A node depends on instances from the previous scheduling cycle. | Dashed lines |
Understanding this distinction before reading a DAG is essential because ancestor and descendant nodes may run on different scheduling cycles. When the number of scheduling cycles or the scheduling time between ancestor and descendant instances differs, preview the dependencies before deploying to verify the configuration is correct. For dependency principles and examples covering complex scenarios, see Principles and samples of scheduling configurations in complex dependency scenarios.
Preview limitations
Before previewing, be aware of the following limitations:
-
The preview shows only the nearest-level ancestor and descendant nodes of the current node.
-
Save all ancestor nodes of the current node before previewing to ensure the dependency graph is accurate.
-
The generated directed acyclic graph (DAG) is for reference only and may differ from the DAG in the production environment.
If you change the scheduling cycle of a node in the production environment, set the Instance Generation Mode parameter to Immediately After Deployment and redeploy the node. Instances already generated and scheduled for the current day may be replaced, and the preview will not reflect the actual scheduling dependencies for those instances. For details, see Configure immediate instance generation for a task.
Preview scheduling dependencies of a node
DataWorks generates a DAG for a node based on its scheduling dependency configuration. Use the DAG to verify that the dependency configuration meets your expectations. Click any node in the DAG to view its details.
Aggregation methods
On the Node Dependency tab, DataWorks provides three aggregation methods for previewing dependencies: Not Aggregate, Aggregate By Workspace, and Aggregate By Owner. Select the method that fits your review needs. For details on all features available in a DAG, see Appendix: Use the features provided in a DAG.
The following figure shows how the three aggregation methods affect the display on the Node Dependency tab.
Preview scheduling dependencies of an instance
If an auto triggered node generates multiple instances on a scheduling day (for example, a node scheduled by hour or minute), DataWorks generates scheduling dependencies and estimated execution times for all instances based on the node's scheduling dependency configuration and time configuration. Check each instance's dependencies and execution time to confirm they match your expectations. Click any instance to view its details.
For nodes scheduled by hour or minute with a time-specific scheduling parameter (for example, $[hh24]), the scheduling parameter of each generated instance is replaced by the actual scheduling time of that instance in its cycle. For more information, see Appendix: Relationships between scheduling parameters and the data timestamp, scheduling time, and actual running time of a node.