Skip to content
Obsevo
Skip to docs navigation

Docs

Control actions

Retry, stop, activate, deactivate — the complete list of writes Obsevo can make, and the constraints that come with them.

Obsevo can act on n8n as well as watch it. Acting from the incident beats finding the run again in another tab.

A monitoring tool holding a write-capable key is a real thing to be careful about, so the constraints are part of the feature rather than an afterthought.

The complete list

retry_execution
stop_execution
activate_workflow
deactivate_workflow

That is all of it. Each maps to a documented endpoint in n8n's public API — Obsevo never touches n8n's internal REST surface.

Nothing in this path reads or transports execution payloads. A retry is executed by n8n against its own stored data; Obsevo asks, n8n does the work.

What has to be true

Control actions are refused, with a specific reason, unless all of these hold:

  1. Monitoring is not paused for the instance.
  2. The instance is in API-key mode. The agent is outbound-only by design, so there is no route back through it into your network. This is the same property that makes agent mode private — see connecting an instance.
  3. A write-scoped n8n API key is configured. Separate from the read key, and opt-in. Obsevo does not quietly reuse a read key that happens to have write permission.
  4. Your role allows it. Control actions require admin. See roles and permissions.
  5. Your plan includes it — Starter and above.

Every action is written to the audit log with the actor, the target and the outcome.

Documentation