Running a workflow

Click Run. Watch each node report its status. Read the result. That's the loop.

A workflow just after a run: Start, Add VAT and Log total each show a green tick and a Completed label, and the header button now reads Back to edit.

How to run

From the editor

Click the run button at the top right of the editor, or press ⌘↩ / Ctrl+Enter.

Mode	Button	What happens
Production	Run Workflow	Real run. Connector nodes call the real services.
Sandbox	Test Run	Dry run. Connector nodes answer from mock data. See Sandbox mode.

If the workflow has more than one trigger, the button reads Choose Trigger and asks which one to start from. You can also hover over a trigger node and click Run in its toolbar.

Running saves your changes first. If the workflow has never been published, Flero asks No active version and offers Publish & Run, which publishes the saved state as version 1 and runs it.

Automatically

Schedules, webhooks, form submissions and connector events start runs on their own, but only when the workflow is active and has an active version. Automatic triggers run the active version, not your latest unsaved edits. See Versions and activation and Triggers.

From the API

You can start a run from your own systems with an API key. See Running workflows through the API.

What you see during a run

Each node shows a status label and a coloured badge:

Label	Colour	Meaning
Pending	Yellow	Waiting to run.
Running...	Blue	Running now.
Completed: duration	Green	Finished successfully, with how long it took.
Failed	Red	The node failed. Timed-out nodes also show as failed.
Waiting for approval	Orange	Paused at an approval step until someone responds.
Skipped / Stopped	Grey	Not reached (for example the other branch of an If), or the run was stopped.

Click any node during or after the run to see its data on the Schema and JSON tabs.

Stopping, pausing and resuming

While a run is in progress, the run button is replaced by:

Stop (red): ends the run. The node that is running is interrupted, no further nodes start, and the run is marked stopped.
Pause (yellow): holds the run. The button changes to Resume (blue) to carry on from where it paused.

When the run finishes, click Back to edit to return to editing. See Rerun, stop & pause.

Past runs

The Production Executions (or Sandbox Executions) list in the header shows this workflow's runs, with status, version, start time, duration and mode.

Click a run to open it read-only on the canvas, with every node's data from that run. Click Back to edit to leave.
Load to Editor (on hover) keeps you in edit mode and fills the nodes with that run's data.
Open log details opens the execution detail page, with its Overview and Output tabs.

See Execution history.

Rerunning

Click Rerun on any finished run (successful, failed or stopped). The rerun uses the same input and the exact version the original run used, even if you've edited the workflow since. A toast confirms "Rerun started". See Rerun, stop & pause.

Handling failures

By default a failing node stops the run. To change that for one node, open Node Settings from the node's More button and pick an Error Mode: retry with backoff, use a fallback value, or continue anyway. You can also set a Timeout. Nodes have no timeout unless you set one. See Configuring nodes and Retries & error handling.

Tips & gotchas

A manual run doesn't arm the triggers. Webhooks and schedules only fire when the workflow is active and has an active version.
Production means real. A run in Production executes real actions: real HTTP calls, real Slack posts, real emails. Use Sandbox while you build.
Pinned nodes don't run. A node with pinned data is skipped in every run, Production included. See Testing while you build.
Invalid credentials fail fast. The node fails at the start with an error you can read in its panel.
Approvals pause the run. The run waits until someone responds. Approval requests appear under Notifications.
Stuck on Running...? Open the run in Logs. A run whose worker stops responding is marked failed after a few minutes. See No worker running and Debugging failures.

Found something out of date? This page lives in the Flero docs content set.