This topic answers common questions about scheduled nodes that do not run as expected in DataWorks.
Only nodes deployed to Operation Center in the production environment are automatically scheduled. Before troubleshooting, confirm that your nodes are published to the production environment. For details, see Publish tasks.
What are the conditions required for a node to run?
A scheduled node runs only when all four of the following conditions are met:
-
Scheduling resources are available.
-
The scheduling time has arrived.
-
All ancestor nodes have run successfully.
-
The node is not frozen.
In Operation Center, auto triggered instances are color-coded by state. Use the table below to identify what each color means, the likely cause, and where to check next.
| Color | State | Likely cause | Where to check |
|---|---|---|---|
| Green | Run successfully | — | — |
| Gray | Not run | An ancestor node has not completed, or the workflow is isolated. | Right-click the instance in the directed acyclic graph (DAG) and select Show Ancestor Nodes. Use the ancestor node analysis and intelligent diagnosis features to trace the root cause. |
| Red | Failed to run | The instance encountered an error during execution. | Review the execution log. |
| Blue | Running | If the instance stays in this state for a long time, scheduling resources may be insufficient. | See Waiting for resources. |
| Yellow | Waiting time | The scheduling time has not arrived. | In the DAG, click Show Details, then open the General tab to check the scheduled time. |
| Yellow | Waiting for resources | The resource group has reached its concurrency limit. | Right-click the instance in the DAG and select Instance Diagnose. In the Resources step, view which nodes are occupying resources. See Waiting for resources. |
| Purple | Freeze | The instance is frozen. Descendant nodes are blocked from running. | In the DAG, click Show Details, then open the Operation Log tab. |
If all nodes in a workflow are gray and the ancestor nodes are not in any of the states above, the dependency between the node and its ancestors may have changed, isolating the workflow. See Scenario: Isolated node.
Why is an auto triggered instance not run after its scheduling time arrives?
The scheduling time and the actual start time of an instance can differ. When an instance does not start after its scheduled time, use the following steps to diagnose the cause.
Use ancestor node analysis to identify the blocking instance, then use intelligent diagnosis to find the root cause.
The following four scenarios cover the most common causes. The examples use a three-layer dependency — the same logic applies to deeper dependency chains.
Scenario 1: Is an ancestor instance still running or failed?
Symptom: The instance is in the Not running state (gray).
Example: The current time is 01:00. Instance A2 is scheduled for 01:00, Instance A3 for 03:00, and Instance B for 00:00. Instance B depends on A2 and A3. Although B's scheduling time has passed, B cannot start until both A2 and A3 complete successfully.
Cause: An instance waits for all ancestor instances to finish before it can run. If any ancestor fails or is still running, the downstream instance stays in the Not running state.
If an ancestor instance stays in the Running state for an extended period:
-
For non-batch synchronization nodes, investigate the execution log to identify the cause.
-
For batch synchronization nodes, the likely causes are resource contention or slow data processing. See How to troubleshoot a batch synchronization node with a long execution duration.
Scenario 2: Has the scheduling time arrived?
Symptom: The instance is in the Waiting time state (yellow). 
Example: The current time is 04:00. A2 is scheduled for 01:00, A3 for 03:00, and B for 05:00. B depends on A2 and A3. Both A2 and A3 have completed, but B's scheduled time has not arrived yet.
Cause: The instance is queued and starts automatically when its scheduled time arrives.
Scenario 3: Are scheduling resources available?
Symptom: The instance is in the Waiting for resources state (yellow). 
Example: The current time is 04:00. A2 is scheduled for 01:00, A3 for 03:00, and B for 00:00. B depends on A2 and A3. Both A2 and A3 have completed, but the resource group running B has reached its concurrency limit.
Cause: The number of nodes running on the resource group in the current workspace has reached the upper limit. The instance is waiting for gateway resources to become available.
To investigate resource usage on an exclusive resource group:
-
In the DataWorks console, click Resource Groups in the left navigation pane. On the Resource Groups page, view running nodes and resource utilization.
-
Right-click Instance B in the DAG and select Instance Diagnose to see which nodes are occupying resources.
Scenario 4: Is the instance frozen?
Symptom: The instance is in the Frozen state (purple). 
Example: The current time is 04:00. A2 is scheduled for 01:00 and is frozen, A3 is scheduled for 03:00, and B is scheduled for 00:00. B depends on A2 and A3. Because A2 is frozen, B is blocked and cannot run.
Cause: An instance cannot run if it is frozen or if any of its ancestor instances are frozen — even if the scheduling time has arrived and all other conditions are met.