2. Click Project List on the left menu to find the target project that needs OPS operation.
3. After selecting the project, click to enter the Data Development Module.
4. Click Task Ops on the left menu.
Task Ops Overview
After the user completes the configuration of computing tasks and the orchestration of task nodes in Data Development > Orchestration Space and submits the schedule, they can manage and operate data workflows and computing tasks in the Task OPS interface. This includes operations like starting, pausing, decommissioning, and deleting workflows, as well as starting, pausing, decommissioning, and deleting computing tasks. Additional features include viewing workflow details, run history, upstream and downstream relationships between tasks, and task execution instances.
Workflow Operation and Maintenance
Workflow List
Select the Workflow List tab on the Task Ops page. The Workflow List page displays data workflows that have been published and entered the maintenance phase. You can view the current status and schedule cycle of workflows in the list, and perform various maintenance operations to ensure efficient, smooth, and reliable data processing flows.
Workflow List Information:
Feature Description:
Information
Description
Workflow name
Display the name and ID of the workflow. Click the
icon to copy the corresponding workflow name or ID. Workflow IDs are used in scenarios like querying logs and troubleshooting in the background. Click the workflow name to go to the detailed information page of the corresponding workflow.
Submission Time: Displays the date and time of the first and most recent maintenance submissions for the current workflow in the format: yyyy-MM-dd HH:mm:ss.
Task Information: Displays the number of compute task nodes involved in the current workflow's orchestration.
Type Distribution: Uses a ring chart to show the proportion of different types of compute task nodes in the workflow.
Cycle Distribution: Uses a ring chart to show the proportion of different scheduling strategies for compute task nodes in the workflow.
Run History: Displays the run history of the workflow (only when all tasks under the workflow are in the same cycle, will the run history be displayed).
DAG Graph: Displays the DAG orchestration diagram of the current workflow and computing nodes, showing types of tasks and node orchestration logic. Moving the cursor over a task node reveals the corresponding task ID, project name, workflow name, owner, task type, task status, task cycle, and scheduling plan.
Folder
Displays the folder directory to which the workflow belongs in the orchestration space, making it easy to identify the workflow's ownership for unified management.
Owner
Displays the name of the person in charge of the workflow, specifying the responsibility for workflow operations. By default, it is the user who created the workflow.
Tasks
Displays the number of compute task nodes involved in the orchestration of the workflow.
Status
Running status of the workflow, allowing users to determine whether the workflow is running as expected based on its status. There are a total of 6 statuses: All Scheduling, All Paused, All Offline, Partially Scheduling, All Unscheduled, All Invalid.
All scheduled: All tasks in the workflow are "scheduled"
All paused: All tasks in the workflow are "paused"
All offline: All tasks in the workflow are "offline"
Partially scheduled: Some tasks in the workflow are "scheduled", while others are "paused" or "offline"
Not scheduled: None of the tasks in the workflow are "scheduled". They can be "offline", "paused", or "invalid"
All invalid: All tasks in the workflow are "invalid", unable to start or pause.
First Submission Time
Displays the date and time when the workflow was first submitted for operations.
Last submission time
Displays the date and time when the workflow was last submitted for operations after the most recent modification.
Workflow Operation and Maintenance:
Feature Description:
Operation
Description
Start
Select the workflow name and use the 'feature' button above to batch start, pause, or offline the selected workflows.
Start: Only effective for tasks with the status "Paused." After starting the workflow, all tasks in the workflow will be in normal operating status and instances will be generated according to the scheduling cycle.
Pause: Only effective for tasks with the status "Scheduling." After pausing the workflow, the tasks in the workflow will not generate new instances.
Offline: Only effective for tasks with the statuses "Scheduling," "Paused," and "Expired." After going offline, the task status will be set to "Offline" and will not generate instances. To re-execute, you need to resubmit in the orchestration space.
Pause
Offline
Start All
For workflows with the running status "All Paused," "Partially Scheduling," or "All Not Scheduled," you can start all nodes that are in the "Paused" status.
Take All Offline
For workflows with the running status "All Scheduling," "All Paused," "Partially Scheduling," "All Not Scheduled," or "All Expired," you can take all nodes in the statuses "Scheduling," "Paused," or "Expired" offline. After going offline, the task status will be set to "Offline" and will not generate instances. To re-execute, you need to resubmit in the orchestration space.
All paused
For workflows with the running status "All Scheduling" or "Partially Scheduling," you can pause all nodes in the "Scheduling" status. After pausing a task, it will not generate new instances. Paused tasks can be restarted to return to the scheduling status.
Data Replenishment
Perform supplementary executions for failed scheduling instances and historically missing instances under the computing tasks, supplementing the missing or failed data processing task cycles. For more details, see Data Supplement.
Modify Owner
The default person in charge of the workflow is the creator of the workflow. You can modify it to other users under the account and designate multiple users as persons in charge.
Deletes a workflow
To delete a workflow in operations, you need to take all the computational tasks in the workflow offline, ensuring the workflow's running status is "All Offline" before deleting the workflow. Workflows deleted from operations will not be removed from the orchestration space and can be resubmitted for operations from the orchestration space later.
Workflow Development
Jump to the corresponding Workflow Canvas in the Orchestration Space, which serves as a shortcut for Workflow Operations and Workflow Editing.
Workflow Filtering:
Operation
Description
Filter by Folder
Support filtering workflows under maintenance by the folder directory to which the workflow belongs in the Orchestration Space. Multi-level directories can be expanded and multi-selection is supported.
Filter by Responsible Person, Workflow Name, Workflow ID
1. First, select a filter Tag, then enter the corresponding filter keyword for the Tag. If there are multiple keywords to be filtered under a single Tag, separate the keywords with a "|".
2. After configuring one Tag, press Enter to select the next Tag and enter the corresponding filter keywords. Repeat these steps to achieve multi-condition combined filtering of OPS workflows.
Workflow Refresh
View the latest status of the workflow list. This is generally used to check the running status of workflows after operations, helping users to get the latest information on maintenance workflows.
Task Operation and Maintenance
Task Ops List
Select the Task List Tab in the Task Operation and Maintenance page. The Task List page displays Compute Task Nodes in published workflows, where you can view the current Scheduling Status, Scheduling Period, and other information. Through a series of feature operations, you can maintain the compute tasks to ensure data processing nodes are stable, effective, and reliable.
Task List Information:
Feature Description:
Information
Description
Task Name/ID
Support copying task names and IDs. Click the task name to jump to the Task Detail Page > Task Lineage.
Owner
Displays the name of the Responsible Person for the compute tasks, specifying the responsibility for task operations. By default, it is the user who created the compute task.
Status
Running status of a computing task, allowing users to determine whether the task is running as expected. There are a total of 5 statuses: Scheduling In Progress, Paused, Offline, Expired.
Scheduling In Progress: The computing task is in a normal "Scheduling In Progress" state, and task instances will be generated according to the schedule.
Paused: The computing task is paused and in a "Paused" state. No new instances will be generated, but existing instances will continue to run.
Offline: The computing task is offline and in an "Offline" state. No new instances will be generated. To reschedule, please go to the Orchestration Space and resubmit.
Expired: The task is in an "Expired" state, indicating that the current time has exceeded the task's configured effective time. Expired tasks cannot be started or paused.
Folder
The folder directory to which the computing task belongs in the Orchestration Space, making it easy to identify the task's ownership for unified management.
Related Workflow
The workflow name to which the computing task belongs in the Orchestration Space, making it easy to identify and manage the task. Clicking the workflow name will navigate to the corresponding workflow’s configuration page.
Task Type
Task Type of the computing task.
Scheduling Type
Scheduling period of the computing task.
Execution Resource Group
Execution Resource Group configured for the computing task.
Last submission time
Displays the date and time when the computing task was last submitted for operations after the most recent modification, in the format: yyyy-MM-dd HH:mm:ss.
First Execution Time
Displays the first time a computing task was run and an instance was generated according to the scheduling period, in the format: yyyy-MM-dd HH:mm:ss.
Creation Time
Displays the date and time when the computing task was first submitted for operations, in the format: yyyy-MM-dd HH:mm:ss.
Task Operation and Maintenance:
Feature Description:
Operation
Description
Start
Divided into batch operations and single task operations. For batch operations, check the box in front of the task name, then use the feature button above; for single task operations, click the action items in the Operation column of the task list.
Start: Only effective for tasks in the "Paused" state. After starting the task, it will be in a normal scheduling state and instances will be generated according to the scheduling cycle.
Pause
Pause: Only effective for tasks in the "Scheduling" state. After pausing the task, no new instances will be generated. You can choose whether to terminate already generated instances during the pause operation.
Offline
Pause: Only effective for tasks in the "Scheduling", "Paused" and "Invalid" states. After taking the task offline, its status will be set to "Offline" and no instances will be generated. To execute again, go to the arrangement space to re-submit.
Data Replenishment
Data Supplement: Completes tasks for failed instances during task scheduling and execution, as well as historical missing instances, supplementing the data processing task cycles. For more details, see Data Supplement
Setting Alarms
Alarm settings: Set alarm monitoring for the task's running status, so that operations personnel can quickly detect and respond to task execution anomalies. For more details, see Task Alarms
Modify Owner
Modify Responsible Person: Change the person in charge for the task to clearly assign responsibility for computing task operations and maintenance. Multiple responsible persons can be added for the same task at one time.
Delete
Delete: Only supports deleting computing tasks in the "Offline" state. You can send a notification of the task deletion to the task responsible person. If multiple computing tasks are selected, other tasks in running states will not be deleted by default! There are two ways to delete:
After deletion, downstream tasks will continue running normally.
After deletion, downstream tasks will fail to run.
View Instance
Jumps to the list interface of the corresponding computing task's generated instances. You can view instance running status, running time, etc., and perform feature operations on the instances.
View details
Jumps to the task detail page, supporting basic task operations and viewing basic task information. For details on the task detail page, see below.
Task Development
Jumps to the corresponding compute node configuration interface in the Orchestration Space. It is a quick channel for Task Operation and Maintenance - Compute Task Node Editing.
Task Filtering:
Operation
Description
Filter by Task Name/ID
Supports fuzzy matching by task name or task ID.
Filter by Folder
Supports filtering tasks in operation and maintenance by the folder name of computing tasks in the orchestration space. Multi-level directories can be expanded and multiple selections are supported.
Filter by Workflow
Supports filtering by the workflow name to which the computing task belongs.
Filter by Task Type/Data Source
Supports filtering by task type. If you select an offline synchronization task, you can filter by source data table and target data table.
Task Details Page
The Task Details page is a collection of task information and operations, including basic information, basic operations, execution status of historical instances, and task lineage. Task operation and maintenance can be performed on this page.
The Task Detail Page contains three tabs: Execution History, Execution Curve, and Task Lineage.
Content
Description
Execution History
Displays the basic information of historical task instances.
Filter: Supports filtering instances by instance data time and start time.
Statistics: Statistics of the total number of instances within the filter range, including the number of instances in each state (waiting for upstream, waiting for event, waiting to run, running, successful, and failed).
Instance List: Displays basic information and operations of instances.
Execution Curve
Display the Execution Time Curve of Task History Instances.
Execution Duration Trend: View the execution duration trend of the last 25, 50, and 100 instances. Users can understand the changes in the execution duration of historical instances, assisting in determining whether there are anomalies in instance execution.
Start/End Time Change Trend: View the change trend of the start and end times for the last 25, 50, and 100 instances.
Task Lineage
Display the lineage information of the current task.
Divided into DAG Pattern and List Mode. For complex lineage relationships and scenarios with many nodes, the List Mode is recommended.
Yes
No
Was this page helpful?