Margo Overview

Fleet Manager is built around Margo, an open specification for managing workloads on fleets of edge devices in industrial automation. Margo defines how a device onboards, reports what it can run, learns what it should be running, and reports back how that’s going.

Who this is for: OT engineers and integrators connecting devices to Fleet Manager.

How a device talks to Fleet Manager

Devices communicate with the SBI Gateway (the Southbound Interface) using the Margo Workload Management API. Fleet Manager implements that API to the letter so that any Margo-compliant device interoperates without custom glue.

The device-side lifecycle has four stages:

1. Onboarding — the device registers and receives its client identity.
2. Capability reporting — the device tells the platform what hardware and interfaces it has.
3. Desired state — the platform tells the device which applications it should run.
4. Deployment status — the device reports the progress of each deployment back to the platform.

See Constraints & Interoperability for the rules the platform enforces to stay Margo-compliant.

Key facts

• Authentication uses RFC 9421 HTTP Message Signatures — there is no mTLS requirement.
• Manifests are JSON, identified by content-addressable sha256: digests.
• Only two deployment profiles are supported: helm.v3 and compose.
• Application deployment YAML uses the API group application.margo.org/v1alpha1.