Skip to main content
After each verified Gateway release, the Gateway brings the managed workload Nodes to the same release, one Node at a time. The fleet family shows that rollout and resumes it after a halt. Fleet rollout explains the rollout set, the order, the steps on each Node, and the catch-up.

​
Commands

Both commands accept --json. JSON output contains only the response.

​
orbit fleet:rollout:status

Show the newest fleet rollout.
The detail shows whether the rollout is on, its status, the commit the Gateway wants the fleet on, and, for a halted rollout, the Node, error code, and message. A table lists each Node in rollout order with its outcome, the step that stopped it, and the CLI and agent versions it runs. The last line names each Node the rollout leaves out, with its reason. --json returns enabled, status, desired_commit, rollout with desired_state and one entry per Node, including its evidence, and excluded.

​
orbit fleet:rollout:resume

Resume the halted rollout. The Gateway runs it in orbit-fleet-converge.service; the command returns at once.
Without --skip, the rollout visits the failed Node again first. The command fails with fleet.rollout_not_halted when no rollout is halted, and with fleet.node_not_in_rollout when --skip names a Node outside it.