Skip to content

Fleet and Releases

This is where you watch every deployment in the field and control what software it runs: the fleet dashboard with heartbeat freshness, maintenance windows, releases and feature flags.

Who can do this:

  • support — fleet dashboard and alerts, maintenance windows.
  • release_manager — releases, maintenance windows and feature flags (releases.manage, maintenance.manage, flags.manage).

The Fleet dashboard (fleetGET /fleet) lists every deployment with its heartbeat freshness — re-derived live, never mocked. A deployment whose heartbeat goes stale surfaces as an alert.

  • Deployment detail (fleet/deployment/:id), plus regions, heartbeats and fleet settings.
  • Fleet alerts (fleet/alerts) — open an alert and Acknowledge it (POST /fleet/alerts/:id/acknowledge). The acknowledgement is audited.

Heartbeats and the on-prem grace/lockdown behaviour are covered in Licensing.

Schedule downtime and tell customers about it:

  1. Open Maintenance (releases/maintenance).
  2. Create a window (releases/maintenance-newPOST /maintenance).
  3. Broadcast it (POST /maintenance/:id/broadcast) so affected tenants are notified. Schedule and broadcast both persist and audit.

Manage the software customers run:

  1. Releases (releases) lists releases; open one for detail.
  2. Create a release (releases/release-newPOST /releases).
  3. Set a tenant’s update channel (PUT /tenants/:id/channel) to control which release line they receive.

Creating a release requires releases.manage and is audited.

Toggle features per the rollout plan:

  1. Open Feature Flags (releases/feature-flags).
  2. Edit a flag (PUT /feature-flags). Changes persist and audit; editing requires flags.manage.