Control plane and node pools are different
The managed AKS control plane contains the Kubernetes API and other control-plane components. Worker nodes live in node pools. An upgrade plan should therefore identify the target Kubernetes version and whether the operation changes the control plane, a node pool, or both.
What happens to workloads?
When nodes are upgraded, AKS can cordon and drain nodes and replace or upgrade the underlying nodes according to the upgrade process and configuration. Workloads may be rescheduled onto other nodes. Proper replica counts and Pod Disruption Budgets help reduce disruption.
Production checklist
- Check Kubernetes version compatibility and the supported upgrade path.
- Review workload replicas and Pod Disruption Budgets.
- Check node pool capacity and autoscaling.
- Validate critical add-ons and ingress components.
- Confirm application health checks and rollback/recovery procedures.
- Perform the change during an approved maintenance window.
Useful verification
kubectl get nodes kubectl get pods -A kubectl get events -A --sort-by=.lastTimestamp
Interview tip: Explain that a node-pool upgrade is primarily about replacing/updating worker capacity, while the control-plane upgrade changes the managed Kubernetes control-plane version. Always verify the exact behavior for the AKS version and upgrade mode in use.