Start a workflow now, outside its schedule, for example after you fix a table or add a new one. Each start creates one run, which you follow in the runs table.
Cancel closes the dialog without starting a run. The Start button reads Running... while the start is sent. The dialog closes as soon as you click Start, so a refusal arrives afterwards as an Error notification with the reason.
After you click Start, a success notification reads Flow scheduled successfully. That message means the run has started. It does not mean the data has loaded.
The run appears in the runs table straight away with Execution Type set to Manual, and the runs list and the timeline refresh to show it. Follow the run in Job runs; to watch its steps, see Monitor and fix a run.
A run is planned when you start it, so its steps are fixed at that moment. A table added to the workflow afterwards waits for the next run.
| Setting | What it does | Default |
|---|---|---|
| Select Date | Sets the run's Current Date, the operation date the run works with. | Today |
| Select Time | Sets the time for that date, down to the second. | Now |
The dialog sets only the run you start. It does not change the workflow's schedule or its settings.
Run is disabled when the workflow has nothing to run yet. Hover over the disabled button to see Can’t run yet and the reason:
In the workflow tab's right-click menu, the item is disabled and reads Run - deploy a table first. While the panel is open, other cards keep their hover actions hidden.
If a start still reaches the server while the workflow has nothing ready, it is refused with Unable to start <name>, at least one table must be deployed.
You do not have the Execute permission for that workflow's type. Ask an administrator for the matching Execute permission, such as Execute Lake Workflow Objects. See Workflow permissions.
Run counts only the tables of this workflow. A deployed table that belongs to another workflow does not count. Move the table to this workflow from the Lake. See Workflows in the Lake.
It sets the run's Current Date, the operation date the run works with. The runs table shows it in the Current Date column. For a normal run, leave it as it is.
Yes. The schedule switch does not stop a run you start by hand, and starting a run does not turn the schedule on.
A run of that workflow is still Running or Running Error. The message reads The <name> cannot start, there is unfinished job(s). Wait for the run to finish, or stop it: Stop is available while a run is Running or Running Error. See Monitor and fix a run.
Another run is preparing or loading the same data source. The message reads Data source "<code>" is being loaded by job <id> (<workflow>). Start again when that run ends.
Yes. Open the table in the Lake and click Synchronize to load that one table. Its run has Run Scope set to Single Node. See Run, stop and troubleshoot loads; to find the run, filter the Run Scope column in Job runs.
Not yet. The run has started. Watch its Status in the runs table until it shows Finished. See Job runs.
The dialog closes as soon as you click Start, so the refusal arrives afterwards as an Error notification with the reason from the server. When the server sends no reason, it reads An unexpected error occurred. Fix the cause, then start the run again. See Troubleshooting workflows.