All Products
Search
Document Center

DataWorks:Abnormal task instance status

Last Updated:Jun 20, 2026

This topic describes how to troubleshoot common task instance issues, such as ungenerated or failed instances.

Task not found in Operation Center

  • Symptom

    The target task does not appear on the Cycle Task page in Operation Center.

  • Cause

    The node has not been deployed to the production environment.

  • Resolution

    Verify that the node was deployed successfully. Check for any deployment failures.

A periodic task generates no instances

  • Symptom

    A task is configured for periodic scheduling, but no instances are generated.

  • Cause

    Periodic scheduling may be disabled for the workspace.

  • Resolution
    On the Workspace Configuration page, go to the Scheduling Settings section and ensure the periodic scheduling switch is on (green).
    Note If required, contact the Alibaba Cloud account holder to enable scheduling for the workspace.

Specific task generates no instances

  • Symptom

    Other nodes in your workflow generate instances as scheduled, but one specific node fails to generate an instance.

  • Causes
    • The node was deployed after the 23:30 cutoff time.
    • The node has no ancestor nodes and has become an isolated node.
    • The node or its ancestor nodes are outside their configured validity period.
    • The node is newly created and its Instance Generation Mode is set to Next Day.
  • Resolutions
    • Check the deployment time. A notice in the Confirm Deployment dialog box on the Task Deployment page explains that for T+1 nodes (next-day generation), changes deployed before 23:30 take effect on the same day, while changes deployed after 23:30 take effect the next day. If your node was deployed after this time, no instance is generated for the current day.
    • Check for isolation. Right-click the Cycle Task DAG to see if the node is isolated and lacks ancestor nodes. Resolve any isolated node alerts immediately.
      Note A node can become isolated if its dependencies change, leaving it without any upstream input. To fix this, redefine its upstream and downstream dependencies and redeploy the node.
    • Check the validity period. Expired tasks do not generate periodic instances. In Operation Center, view the instance Properties and check the Scheduling Effective Period field. If the date range is expired, such as 1970-01-01 00:00:00 ~ 1970-02-01 00:00:00, the task has expired. Go to Scheduling Settings in DataStudio to update the validity period.

Instances are generated but do not run

  • Symptom

    An instance is generated for a Cycle Task but remains in the To be run (gray) state and does not execute.

  • Cause

    An ancestor node has not completed successfully, which blocks all downstream execution.

  • Resolution
    In the DAG, right-click the gray instance and expand its ancestor nodes to check for any in the Running, Failed, Waiting, or Frozen state.
    Note For more troubleshooting steps, see Tasks not running.
    Ancestor node status
    • Purple (Frozen)

      The instance is frozen and will not run, blocking all descendant nodes. Check the Operation Log on the instance details page to determine who froze the instance and why.

    • Yellow (Waiting)

      The instance is waiting for its scheduled time to arrive. Open the instance details to view the scheduled time.

      Alternatively, it may be waiting for scheduling resources, which will be indicated in the logs. This occurs when the project's task concurrency limit is reached. To identify which tasks are using resources, right-click the instance and select Instance Diagnose.

    • Gray (To be run)

      The instance is waiting for its ancestor nodes to complete. To trace dependencies, continue right-clicking each gray instance and expanding its ancestor nodes.

      If the entire dependency chain is gray, the starting node is likely an isolated node that lacks the dependencies needed to trigger it.

    To accelerate troubleshooting, use the built-in diagnostic tools. In the Operation Center DAG, right-click a node and select Show Ancestor Nodes > upstream analysis to analyze dependencies, or click Instance Diagnose for an automated check. You can also select Show Descendant Nodes to expand one to six levels and view downstream dependencies.
    Note
    • A task can run only when all conditions are met: its scheduled time has arrived, scheduling resources are available, all ancestor nodes have run successfully, and the task is not frozen.
    • If an entire workflow appears gray without any of the other statuses mentioned, it is likely that a change in dependencies has isolated the entire workflow.

Prerequisites for a task to run

For a task instance to run, all the following conditions must be met:
  • The scheduled time has arrived. An instance waiting for its scheduled time has a Waiting (yellow) status.
  • All ancestor nodes have completed successfully. A successfully run node has a Succeeded (green) status. Check the status of ancestor nodes in the DAG within Operation Center.
  • Sufficient scheduling resources are available. An instance waiting for resources has a Waiting (yellow) status. The logs will indicate the task is waiting for gateway resources because the project's concurrency limit is reached.
  • The instance is not frozen. A frozen instance has a Frozen (purple) status and will not run.