A best practice is a “fast replacement” strategy: maintain a small pool of pre-imaged spares for swap-and-go recovery. This keeps downtime low and support predictable.