FruxonDocs
Agent NetworkParticipants

Deletes a periodic sync policy — the sync stops, its past runs stay in the history.

To pause instead, save the policy with `enabled=false`: that keeps the cadence and the enrollment decision written down, where deleting throws both away. Deleting never touches the contacts a past run imported or the access it granted; those are reversed per run via `POST directory/importRuns/{importRun}:undoEnrollment`.

DELETE
/v1/tenants/{tenant}/directory/syncSchedules/{syncSchedule}
AuthorizationBearer <token>

JWT Authorization header using the Bearer scheme. Enter 'Bearer' [space] and then your token.

In: header

Path Parameters

syncSchedule*string

The schedule id.

Formatuuid
tenant*string

The tenant identifier

Response Body

curl -X DELETE "https://api.fruxon.com/v1/tenants/string/directory/syncSchedules/497f6eca-6276-4993-bfeb-53cbbbba6f08"
Empty
Empty
Empty
{
  "type": "string",
  "title": "string",
  "status": 0,
  "detail": "string",
  "instance": "string",
  "property1": null,
  "property2": null
}

Lists the workspace's periodic sync policies, newest first.

Each entry carries the policy plus the bookkeeping that answers whether it is actually running: when it last enqueued, which run that was, when it is next due, and the reason the last tick failed if it did. A policy whose `lastError` is set has stopped producing runs and needs attention — that is the failure a periodic job hides best.

Creates or replaces the periodic sync policy for one integration config.

There is at most one policy per `(provider, integrationConfigId)` — the config a sync sweeps is the policy's identity — so saving twice for the same config replaces the policy rather than adding a rival one. The whole policy is sent each time: the enrollment half has to be stateable as "off", and re-confirming it on every save is the right amount of friction for the field that decides whether a timer grants people access to an application. The cadence is an interval measured from the last enqueue, bounded to between 15 minutes and 90 days. A freshly saved policy waits one full interval before its first run, so saving is never a way to start a sweep — use `POST directory/import` for that. Re-saving keeps the existing clock, so widening a cadence neither double-fires nor skips a cycle. Omitting `targetNetworkId` — the default — makes the periodic run Directory-only: contacts are created and refreshed, and nobody is granted access. Enrollment on a timer is deliberately an explicit choice, because the compensating control for a bulk sweep skipping the per-contact approve gate is an operator reading the dry-run report, and an unattended run has no one reading it. With a target set, `agentMode=OFF_HUMAN_ONLY` stages new contacts (grant created, agent silent) instead of putting the agent live for everyone the sweep finds.