A system that automatically deploys, scales, restarts, and networks containers across a cluster of machines, so you don't manage each one by hand.
Example: Kubernetes automatically restarting a crashed container and rerouting traffic away from it without a human intervening.
In practice: The operational complexity is real — most teams that outgrow a single server benefit from it, but running your own cluster is a significant commitment most small teams offload to a managed service instead.