DataWorks:View auto triggered instances

Intelligent search mode

The intelligent search feature for auto triggered tasks lets you enter search content directly. The system automatically parses your query and filters the instance list.

1. Perform an intelligent search.

   Click the Intelligent Search button in the filter box. In the dialog box that appears, enter the content you want to search for, such as Sort by instance type, and press Enter. The system automatically matches and displays the relevant instances.

2. Save the new view.

   To save the search conditions for later use, click Unsaved View > Save As New View below the search bar. In the Save View dialog box, specify a custom View Name and click Save. You can then find and use this new view in the view search bar to search for instances.

   Note

   If you no longer want to use this new view, find its name in the view search bar. Hover the mouse pointer over the view name, click the ... button on the right, and select Edit or Delete.

3. Close intelligent search.

   To perform an exact search using filter conditions, press the Esc key or click the Close Intelligent Search button in the instance toolbar to exit the intelligent search mode.

Condition-based filtering mode

The condition-based filtering feature lets you accurately filter the instance list by specifying filter conditions.

1. Simple condition filtering.

   In the toolbar, you can select multiple filter conditions, such as Task Name/Task ID/Instance ID, Schedule Resource Group, Alerts Generated In The Last 24 Hours, and Waiting For Resources For A Long Time, to filter the instance list.

2. Complex condition filtering.

   Click the Filter button in the filter box. You can then combine multiple conditions, such as Task Name/Task ID/Instance ID, Schedule Resource Group, Runtime, and Computing Resource Name, to accurately find the required instances.

Single instance operations

To perform an operation on an auto triggered instance, find the instance in the list and use the corresponding feature in the Actions column. The available features are described as follows:

Batch instance operations

To perform batch operations on auto triggered instances, select the required instances in the list. Then, at the bottom of the list, perform batch operations such as Stop, Rerun, Set To Success, Modify Resource Group, Pause (Freeze), or Resume (Unfreeze).

DAG panel features

On the instance DAG details page, you can use the following features in the DAG panel to aggregate nodes, analyze downstream dependencies, and adjust the DAG display style.

DAG operations

On the instance DAG details page, right-click an instance in the flow to view its upstream and downstream relationships, code details, and other related information. The specific operations are as follows:

• Expand Parent Nodes: View the upstream tasks of the current node to understand which nodes affect its data output. You can expand parent nodes by level, up to six levels at a time.

• Expand Child Nodes: View the downstream tasks of the current node to understand which nodes are affected by its data output. You can expand child nodes by level, up to six levels at a time.

• View Runtime Log: After a task starts to run, you can view its detailed execution process in the runtime log. For a description of the core parameters in the log, see Appendix: Description of runtime log parameters.

• Run Diagnostics: Diagnose upstream dependencies, time scheduling, scheduling resources, and task running status.

• View Code: Confirm the current code of the node in the production environment. If the code is not as expected, check whether the latest version of the node has been successfully published.

• Edit Node: Go to the Data Development page and open the current node.

• View Lineage: View the data lineage of the current instance.

• View More Details: View the basic properties, operation logs, and task code of the instance.

• View Auto Triggered Task: View information about the auto triggered task to which the current instance belongs.

• Go to Task 360: In the Data Governance center, you can obtain a panoramic view of a task from multiple dimensions, such as its associated baselines and instance running status, and perform task administration. For more information, see Get a panoramic view of a task.

• Stop: Stop a task that does not need to run. After being stopped, the task exits with a failed status. You can stop only instances that are in the Waiting For Time, Waiting For Resources, or Running state.

• Rerun: Rerun a task that is in the Successful or Failed state. After the task runs successfully, it can trigger the scheduling of downstream tasks that have not run. This is often used to handle error nodes and missed nodes.

• Rerun Descendant Nodes: Rerun the descendant nodes of a task that is in the Successful or Failed state. You can select the downstream tasks to rerun. After a task runs successfully, it can trigger the scheduling of downstream tasks that have not run. This is often used for data fixing.

• Set To Success: Set the status of a failed task to successful. Use this feature when a task fails, but you do not want it to block its downstream tasks. This is often used to handle error nodes.

• Resume: Resume a failed task from the point of failure. If a task contains multiple SQL statements, it can be resumed from the specific SQL statement that failed.

  Note

  This operation is supported only for MaxCompute SQL tasks.

• Trigger DQC Check: If the task is configured with data quality rules, you can trigger a check of those rules.

• Emergency Operation: An emergency operation is effective only for the current run of the current node.

  • Remove Dependency: Remove a dependency for a specified task. Use this feature to remove the dependency relationship of the current node. This is often used to urgently remove an upstream dependency when an upstream task fails and has no data relationship with the current instance, allowing the current task to run.

    Note

    Confirm whether this operation affects data based on the task code and data lineage.

  • Modify Priority: The priority of an instance is inherited from the priority of the baseline to which the instance belongs. You can reset the priority here as needed. A larger value indicates a higher priority.

  • Force Rerun: Forcibly rerun the current node. This operation is supported for auto triggered instances that are successful, failed, or have not run. It is often used for data fixing.

  • Force Rerun Descendant Nodes: Rerun data for the data timestamps of yesterday and the day before yesterday. This operation is supported only for auto triggered instances that are successful or failed. It is often used for data fixing. For more information, see Appendix: Force rerun descendant nodes.

    Note

    Only workspace administrators, tenant administrators, and Alibaba Cloud accounts can initiate this operation.

  • Clone Instance: Create a new instance (clone instance) with the same configuration based on the current instance that is in the Running state (host instance). The new instance is named in the format dw_clone_Node name.

    Note

    • You can clone only ODPS SQL node instances. Each instance can be cloned only once.

    • Execution logic for host and clone instances:

      • Both the host and clone instances are in the running state. If the host instance succeeds first, the clone instance is stopped. If the clone instance succeeds first, the host instance is stopped and its status is set to successful.

      • If a downstream task of the current task has a clone instance, rerunning the downstream tasks does not trigger the execution of the clone instance.

• Pause (Freeze): Freeze an instance if it and its downstream instances do not need to run.

  Important

  • A frozen upstream instance blocks its downstream instances. Use this operation with caution.

  • Do not operate on the projectname_root node. This is the root node of the workspace. All instances of auto triggered tasks depend on this node. If you freeze this node, the instances cannot run.

  • You cannot freeze an instance that is in the Waiting For Resources, Waiting For Time, or Running state (for example, the node code is running or data quality is being checked).

• Resume (Unfreeze): Unfreeze a frozen instance.

  • If the instance has not run yet, it runs automatically after its upstream tasks are complete.

  • If all upstream tasks are complete, the task is directly set to failed. You must manually rerun the instance for it to run normally.

  Note

  The unfreeze operation affects only the current instance. If the auto triggered task is still frozen, the instance generated the next day will also be frozen.

View instance details

In the instance DAG, left-click a specific instance. In the window that appears, click View Log, or Expand Details to view detailed information such as Properties, Runtime Log, Operation Log, and Code.