Norsk Manager
Norsk Manager runs Norsk media systems across a fleet of machines instead of one. You describe what should run, where it is allowed to run and when; the Manager finds suitable capacity, launches the work, watches it, moves it when a worker is lost or drained, and reclaims the capacity when the work ends. norsk-ctl manages a single host. Norsk Manager is what you reach for when one host is not enough, or when hosts should come and go with the schedule.
The manager host runs the control plane. The machines doing the media work are worker nodes. Workers can be local servers that you own and register, cloud nodes that the Manager provisions and terminates, or a mixture of both. In either case, the Manager owns the fleet: it connects to each worker, keeps the media containers running, keeps new work off a cordoned node and empties a node when you drain it.
Everything an operator does goes through norsk-mgr: one binary that provides the HTTP API, web UI and CLI, with the same operations in all three. For example, you can drain a node from the Nodes page, with norsk-mgr nodes drain, or with POST /api/v1/nodes/{id}/drain. It runs on the manager host and is accessed from elsewhere with an API token.
Concepts is the vocabulary the rest of this manual uses — products and templates, productions and runtimes, pools and nodes, and what the Manager does when a worker goes away.
The pages under Guides are captured rather than merely illustrated: each walkthrough is a test that ran against a real deployment, and every command and screenshot on it is that run’s output, dated at the foot of the page.