Workflows and Executions

Workflows

A workflow is a collection of nodes connected together to automate a process. Each workflow has a single entry point (a trigger node) and can branch, merge, and fan out through any number of action and logic nodes. Workflows are stored as JSON and can be exported, imported, versioned, and shared.

Key properties of a workflow:

Property	Description
Name	Human-readable label shown in the editor and execution list
ID	Unique numeric identifier assigned on creation
Active	Whether the workflow listens for trigger events in production
Tags	Optional labels for organizing workflows
Settings	Per-workflow overrides for timezone, error workflow, retry behavior, and execution timeout

Executions

An execution is a single run of a workflow from start to finish. Every time a workflow is triggered, n8n creates an execution record that captures the input data, output data, and status for every node that ran.

Execution States

State	Meaning
Waiting	Execution is paused, waiting for an external event (e.g., a Wait node or webhook response)
Running	Execution is actively processing nodes
Success	All nodes completed without error
Error	At least one node failed and no error handling caught it

Manual vs Production Executions

• Manual execution -- triggered by clicking "Test Workflow" or "Test Step" in the editor. Data is always shown in the UI. The workflow does not need to be active.
• Production execution -- triggered automatically by a trigger node while the workflow is set to Active. Execution data is saved according to the instance or workflow-level retention settings.

Note

Manual executions always use the test webhook URL, while production executions use the production webhook URL. These are different endpoints.

Execution Data and Logs

Execution data includes the full input and output JSON for every node. Retention is controlled at two levels:

1. Instance-level -- set via environment variables (EXECUTIONS_DATA_PRUNE, EXECUTIONS_DATA_MAX_AGE).
2. Workflow-level -- override in the workflow settings panel under "Save Execution Data."

Tip

On busy instances, prune old execution data aggressively to keep database size manageable. Store only failed executions if storage is a concern.

Related tips: Workflow Architecture, Testing & Debugging

See Also

• Triggers
• Nodes
• Error Handling & Retries
• Workflow Architecture Tips