Operation Action
An Operation action runs one of the platform's datastore operations for you. When the Flow's trigger activates, the action starts its operation with the configuration saved on the node, so routine work such as refreshing schemas, profiling tables, scanning for anomalies, and exporting results happens without anyone opening the datastore. Five types are available: Sync, Profile, Scan, Export, and Materialize.
Why Use an Operation Action
Use an Operation action whenever datastore work should happen on its own, started by an event or a schedule, instead of by someone remembering to run it. Common cases:
- Chain a pipeline behind one trigger: Run a Sync, then a Profile, then a Scan in order. A downstream action only starts when its parent succeeds, so a failed refresh skips the steps that would measure stale data.
- Keep reporting fresh: Run an Export right after a Scan, so the enrichment destination always carries the latest anomalies, quality checks, or field profiles.
- Automate the routine: Refresh and validate a datastore nightly with a Schedule trigger, with no one starting operations by hand.
- React to other operations: With an Operation Completed trigger, start the next stage the moment the previous one succeeds, instead of guessing a safe gap between separate schedules.
The Settings Panel
Picking an operation type (Sync, Profile, Scan, Export, or Materialize) opens its settings panel on the right-hand side of the canvas. Each panel carries the same configuration options as running that operation directly on the datastore, including source datastore, table or file selection, read settings, and the options specific to the operation type.
Initialized datastores only
Only initialized datastores appear in the Datastore dropdown, across all five operation types. A source datastore becomes initialized after successful Sync, Profile, and Scan runs, while an enrichment datastore needs only Sync and Profile. Export and Materialize narrow the list further, offering only source datastores that also have an enrichment destination connected. If a datastore is missing from the dropdown, run those operations on it first and check its enrichment destination.
Save, Then Publish
Operation actions follow the Flow's draft-then-publish model. Clicking Save on the panel stores the configuration as an unpublished change. The node shows a CHANGED badge and the You have unpublished changes banner appears on the canvas. The action only takes effect once you click Publish. See How a Flow Works for the full lifecycle, including canceling unpublished changes and publish validation.
The 10-Action limit
A Flow can contain at most 10 Operation actions. Once the limit is reached, the Operation category is disabled in the List of Actions panel (and the Clone option is unavailable on Operation actions) until one of the existing Operation actions is removed. The other action categories are not subject to this cap.
Node Summaries
Action nodes display inline summaries with key details of their configuration, so you can tell them apart without opening each panel. Operation nodes show the name of the datastore they target, and Export nodes also display the asset type in their titles (for example, "Export Anomalies") to identify the exported content at a glance.
See Also
-
Anomaly
Archive or delete inactive anomalies automatically.
-
Notifications
The dispatch pipeline and the five channels: In App, Email, Slack, Microsoft Teams, and PagerDuty.
-
Message
The message you write, its formatting, and the dynamic tokens available per trigger type.
-
Workflow
Hand the event to an external system through a Webhook, an HTTP action, or n8n.
-
Ticketing
Create or update tickets in your issue tracker from anomaly events.
-
Examples
Real scenarios showing what each action category does inside a working Flow.
-
Best Practices
Guidelines per category for actions that do their work quietly and fail loudly.
-
How It Works
The action node, the five categories in detail, how actions run, and the chaining rules and limits.