Workflows

Overview

Workflows allow users to invoke multiple related functions in a single operation. Instead of running each function separately, users can execute multiple jobs through a single workflow execution, simplifying execution and job management.

Accessing Workflows

To create a workflow:

1. Navigate to Workflows from the left navigation menu.

2. Select Create Workflow.

3. Enter a workflow name and optional description.

Creating a Workflow

The workflow editor allows users to define multiple tasks and their execution flow.

Each task can invoke a deployed function.

Function Selection

The Function Name field displays a list of successfully deployed functions available in the current project.

Users can select the function to be executed for each workflow step.

Configuration

For each workflow step, users can configure:

· Function Name

· Provider

· Device

· Number of Shots

· Input Parameters

After updating any field, click Apply to Code to write the configuration back to the corresponding task definition in the workflow code.

Workflow code and step configuration remain synchronized.

· Changes made in the Steps panel can be written to the code using Apply to Code.

· Saved task configurations are restored automatically when the workflow is reopened.

· Users may switch between code editing and visual configuration without recreating workflow settings.

This synchronization simplifies workflow maintenance while preserving full access to the underlying workflow code.

Editing and Saving

Users can modify the workflow code or step configuration at any time.

To save the workflow:

1. Update workflow steps as needed.

2. Click Apply to Code to synchronize configuration changes.

3. Click Save Workflow.

The workflow is stored and becomes available for execution.

Opening an Existing Workflow

To edit an existing workflow:

1. Navigate to the Workflows page.

2. Select a workflow from the workflow list.

3. The workflow code is loaded into the editor.

4. Task configurations are automatically displayed in the Steps panel.

5. Modify the code or configuration as required.

6. Save the workflow to persist changes.

Running a Workflow

After a workflow has been saved, users can manually trigger a new workflow execution.

To run a workflow:

1. Open the workflow details page.

2. Click Run Workflow in the upper-right corner.

Run Workflow Dialog

In the Run Workflow dialog:

· Select the workflow version to execute.

· Optionally provide input parameters in JSON format.

· Enter {} if no parameters are required.

Click Run to start the workflow execution.

A new workflow run is created and added to the workflow execution history.

Workflow Run History

The lower section of the workflow details page contains the Run History table.

Each row is one execution of the workflow:

Column	Meaning
Run ID	Identifier of the run, for example `run-14`
Created By	The user who triggered the run
Status	`Success`, `Running`, `Failed`, `Cancelled` or `Pending`
Duration	How long the run took
Created At	When the run was queued
Actions	View opens the run details; the restart icon re-runs the execution and is disabled unless the run failed

The row itself is not a link — use the View action to open the run details.

◆The Run Info section displays workflow execution metadata, including:

· Created By

· Started At

· Finished At

· Duration

· Workflow Version

· Prefect Run ID

This information helps users identify and troubleshoot workflow executions.

◆The Parameters section shows the runtime parameters supplied when the workflow was executed.

Example:

This helps users verify the input values used during execution.

◆The Workflow DAG section provides a visual representation of workflow execution status and task dependencies.

Each task is displayed with its execution result:

Success

– Task completed successfully.
Running

– Task is currently executing.
Failed

– Task execution failed.
Cancelled

– Task execution was cancelled.
Pending

– Task is waiting to run.

In the example shown:

simulate-a57

completed successfully.
bell_state-09b

failed during execution.

The DAG visualization helps users quickly identify task execution order, duration, dependencies, and failure points within the workflow.

◆The Step Details section lists the individual job steps the run went through, so you can see how far a run got and exactly where it stopped:

Column	Meaning
Step	The step name from the flow, for example `simulate-05c`
Step Status	The outcome of that step
Started At / Finished At	Step timing
Job ID	The job the step created. It is populated only once that job exists, so a step that never started shows no job
Job Name	The function the job ran
Job Status	The status of that job, for example `Done`

Selecting a step shows its Function Config. The fields depend on the kind of job:

Quantum — Function Name, Provider, Device, Shots
HPC — Function Name, Provider, CPU, Memory

◆The Logs section displays workflow execution logs in real time and after execution completes.

Logs may include:

Workflow submission events
Deployment initialization
Task execution progress
Infrastructure events
Error messages and stack traces
Workflow completion status

Users can filter logs by severity level, sort log entries, and review detailed execution information for troubleshooting and monitoring workflow behavior.

Restarting a failed run

A run that ended in Failed can be restarted. A successful, cancelled or in-flight run cannot — the control is disabled for those.

Start the restart either from the restart icon in the Actions column of Run History, or from the Restart button on the run's own details page.
Confirm the dialog "Are you sure you want to restart this run?".
The run is reset and its status becomes Running. You can follow the progress on the same page.

Cancelling the dialog changes nothing and the run stays Failed. If the restart request itself fails, an error message is shown and the run remains Failed.

⚠️ A restart re-runs the whole workflow, not just the failed step. The run is re-executed from the first step, using the same input as the original run. Steps that had already succeeded run again and consume quota again. See Project Quota.