Ops operations are core actions for troubleshooting and daily Ops. This document provides a unified introduction to the Ops operations supported for workflows and tasks in different states.
Overview
Ops operations are divided into the following categories:
|
Running | Workflow | Immediately trigger a workflow run. |
Advanced run | Workflow | Customize parameters and execution task scope |
Rerun | Workflow Run / Task Run | Rerun completed instances. |
Terminate | Workflow Run / Task Run | Terminate running or waiting instances. |
Deleting an alarm policy | Workflow Run / Workflow | Delete run history or workflow. |
Terminate all | All Runs of the Workflow | Terminate all running instances of the current workflow. |
Terminate all waiting runs. | All Waiting Runs of the Workflow | Terminate only runs in waiting state. |
Prerequisites
You have Run (rerun or terminate) or Manage (delete) permission on the target workflow.
The operating account has the resource access permissions corresponding to the workflow Run As account.
Operation Steps
Step 1: Run and Advanced Run
Running
|
Run at the End of Workflow List Row | Directly trigger a run using the saved parameters. |
Run in the Workflow Details Page Action Bar | Same as above. |
Restart Run in the Workflow Details Page Action Bar (Continuous Run Mode) | Terminate the current run and restart it; subsequent scheduling continues in continuous run mode. |
After the operation succeeds, a message is displayed: Workflow "{workflow name}" ran successfully. View details.
Advanced Run
|
Advanced Run at the End of Workflow List Row | Directly trigger a run using the saved parameters. |
Advanced Run in the Workflow Details Page Action Bar | Same as above. |
1. Click Advanced Run to open the dialog box.
2. DAG diagram on the left: Click nodes to select or deselect them, which determines the tasks to be executed in this run. All nodes are selected by default.
3. Parameter configuration on the right:
Configured parameters of the workflow are displayed by default. The Key cannot be modified, while the Value can be modified.
You can click + Add Parameter to add temporary parameters.
Temporary parameters take effect only for the current run.
4. Click Run to trigger the run.
Step 2: Rerun
Rerun is used to re-execute a specific run, including at the task granularity. The rerun target determines the behavior:
Rerunning a Workflow Run
1. On the workflow run list or run details page, click Rerun.
2. The pop-up window is similar to Advanced Run:
Failed nodes are selected by default in the DAG diagram.
Parameters are filled by default with the values used for the current run and can be adjusted.
3. After Run is clicked, a new execution starts.
Rerunning a Task Run
1. On the task run details page, click Rerun.
2. The pop-up window is similar to rerunning a workflow, but only the current task is selected.
Rerun Limitations
Scenario | Action |
Workflow run not in final state | A single task cannot be rerun (button disabled). |
Workflow dependency changed. | The rerun button is disabled, and a tooltip on hover indicates that task dependencies in the workflow have changed and rerun is not supported. |
Internal loop node in For Each | Rerunning a single iteration is not supported. Rerun the entire For Each node. |
Rerun Parameters and Code
Code: Use the latest version, which is the code most recently saved for the task.
Configuration (except dependencies): Use the latest configuration.
Dependencies: Rerun is disabled when dependencies change.
If alarms are configured for the task, rerunning the task will also trigger alarms.
Step 3: Terminate
Terminate is used to stop instances that are running or waiting.
Termination Rules
Status | Terminable or Not |
Pending / Running / Retrying after failure | ✅ |
Succeeded / Failed / Skipped | ❌ |
Terminating | ❌No (termination initiated) |
Workflow-Level Termination
Operation | Description |
Terminate all | Terminate all runs of the current workflow (including running and waiting). |
Terminate all waiting runs. | Terminate only runs in waiting state. |
Confirmation dialog: Confirm to terminate all executions/all waiting runs of workflow "{workflow name}".
Task-Level Termination
Operation | Description |
Terminate | Terminate the current task run (including running and waiting). |
Confirmation dialog: Are you sure you want to terminate the current task?
Step 4: Delete
Deleting a Workflow Run
1. Deletion is allowed only after the run reaches a terminal state. The button is disabled when the run is not in a terminal state.
2. Confirmation dialog: Confirm to delete workflow run {run ID}. After deletion, the workflow run record cannot be recovered.
Deleting a Workflow
1. On the workflow details page or at the end of a list row, click Delete.
2. Confirmation dialog: Confirm to delete workflow "{workflow name}". After deletion, the workflow and its run records cannot be recovered.
Warning:
Deletion cannot be undone. Deleting a workflow will also delete all run records under that workflow.
Step 5: Common Operation Combinations
Scenario | Recommended Action |
Rerun after modifying code following task failure. | Modify the code → rerun the failed task in the workflow run details. |
Rerun from scratch due to configuration error. | In the workflow run list, click Rerun and select all tasks. |
Task holding resources for a long time | Terminate → troubleshoot resource issues → rerun. |
Scheduling produces excessive invalid runs. | Pause scheduling → delete run records → start scheduling after the fix. |
State Transition Matrix
Workflow Run Operation Matrix
Status | Rerun | Terminate | Deleting an alarm policy |
Pending | ❌ | ✅ | ❌ |
Running | ❌ | ✅ | ❌ |
Succeeded / Failed / Skipped | ✅ | ❌ | ✅ |
Terminating | ❌ | ❌ | ❌ |
Task Run Operation Matrix
Status | Rerun | Terminate |
Waiting for Workflow Concurrency / Waiting for Upstream / Waiting for Task Concurrency / Waiting for Resources | ❌ | ✅ |
Running | ❌ | ✅ |
Retry upon Failure | ❌ | ✅ |
Terminating | ❌ | ❌ |
Succeeded / Failed / Upstream Failed / Skipped / Excluded | ✅ | ❌ |
Failure and Timeout Policy Interaction
For details about the retry and timeout policies of task runs, see "Failure and Timeout Policy" in General Task Configuration. During a rerun, the retry policy remains effective, and the timeout threshold is recalculated from the start time of the rerun. Use Limits
Before deleting a workflow, you must terminate all its runs. Workflows that are running or waiting cannot be deleted.
The For Each inner loop node does not support a single rerun.
FAQs
Q1: How do I select "Failed tasks and their upstream tasks" during a rerun?
In the DAG diagram of the rerun dialog, hold down Shift and click nodes to select multiple nodes, or use the box selection tool to select a chain at once.
If the rerun option supports "Rerun instance scope = Current instance and upstream instances" (in some scenarios), you can switch it in the dialog.
Q2: Can a workflow be recovered after deletion?
No. A workflow and its run records cannot be recovered immediately after deletion. Make sure there are no business dependencies before you proceed.</0>
Q3: If upstream task parameters are modified during a rerun, will downstream references use the new parameters?
Do not rerun upstream tasks: downstream tasks use the parameters from the last successful run.
Rerun includes upstream tasks: downstream tasks use the parameters from the latest upstream results.
Q4: What is the difference between "Terminate All" and "Terminate"?
Terminate: An operation in the list row of a run that terminates only that run.
Terminate All: An operation on the workflow details page that terminates all runs under the workflow, including running and waiting runs.
References