Restore a sandbox from a snapshot, potentially on a different worker.
snapshot_id or the best available), picks an
eligible orchestrator node — always excluding the originating node — forwards
the restore to that orchestrator, and updates sandbox-manager state with the
new node assignment.
Unlike resume, restore works on sandboxes in
any state as long as at least one snapshot exists, and always lands the sandbox
on a fresh worker.
sbx-<8 chars>.pause → periodic → manual).guest_ip and envd_port for the restored VM.
| Status | Cause |
|---|---|
401 | Missing or invalid API key |
404 | Sandbox not found, or no snapshot available |
502 | Orchestrator unreachable or restore failed |
503 | No eligible orchestrator node available |