Running a workflow
Press Run. Watch the canvas colour. Read the result. That's the loop.
🎬 Gif needed:
editor__run-canvas-colours.gif, clicking Run, then the nodes turn yellow → green in sequence, ending with the bottom chrome showing "Completed". ~5 seconds.
How to run
From the toolbar
Click Run in the top toolbar. This is a manual trigger, Flero invokes the trigger node with empty input and lets the workflow proceed.
Keyboard shortcut: ⌘↩ / Ctrl+Enter.
From the trigger node
Some trigger nodes also accept inline payloads. For example, the Webhook Trigger has a Test webhook button in its inspector that lets you paste a JSON body and headers and fire a real run with them. Schedule and Form triggers have similar "fire now with this data" buttons.
From the API
Each workflow has an API endpoint you can POST to. Find the URL in Workflow settings → API trigger. You'll need an API key (see API keys).
What you see during a run
The canvas animates while the workflow runs:
| Colour | Meaning |
|---|---|
| Yellow / pulsing | The node is currently executing. |
| Green | The node finished successfully. |
| Red | The node failed. |
| Grey | The node was not reached (e.g. it's on the false branch of an If that took true). |
| Blue / paused | The node is waiting for input (typically an Approval). |
The bottom execution chrome shows the overall state and the elapsed time. Click any node mid-run to see its (partial) inspector output.
When a run completes
The chrome flips to Completed (green) or Failed (red). The most recent five runs appear in the dropdown beside it, pick any to load that run's data into the canvas snapshot.
The full result is available on the execution detail page, click the execution chip to open it in a new tab.
Stopping a run
Click Stop in the bottom chrome at any time. The currently executing node is allowed to finish (Flero doesn't kill mid-node, that would leave external side effects in an inconsistent state), then the run is marked cancelled and no further nodes execute.
Workflow-level options that affect runs
Open Workflow settings (gear icon in toolbar) to control how runs behave:
| Option | What it controls |
|---|---|
| Execution timeout | Whole-workflow ceiling. Default 300 s. Triggers a timeout status if exceeded. |
| Concurrency | Maximum simultaneous executions of this workflow. Higher = more throughput; too high = downstream service overload. |
| Error policy | On any node failure: fail workflow (default), continue (workflow ends in success even if a node errored), or route to error path (use error-handling nodes). |
| Save successful runs | Whether to persist execution history for successful runs. Failed runs are always saved. |
| Sandbox by default | Force every run into sandbox mode, useful for development workflows. |
Tips & gotchas
- A manual Run does not arm the trigger. Webhook URLs are still inert unless the workflow status is
active. Manual runs are for testing the body of the workflow. - Test mode vs real mode. A manual run executes the real action nodes (real HTTP calls, real Slack posts). Use sandbox mode to mock them.
- If a credential is invalid, the node fails fast at the start of its execution, read the error in the inspector.
- Stuck on "Running…" for a long time? Open
/logs/<id>for the full execution detail, the worker may have crashed or be queued behind other work. See Debugging failures. - Approvals pause the workflow, the status will be
waitingForInputuntil someone responds. Approval requests appear in/approvals.
Related
- Testing a node, single-node iteration
- Execution lifecycle
- Debugging failures
Found something out of date? This page lives in the Flero docs content set.