Changes to Drain a node: mark it as "draining" so no new replicas are scheduled on it, and trigger redeployment of all affected environments so their containers are rescheduled to healthy nodes.