How Should an AKS Cluster Be Configured for Zone and Maintenance Resiliency?

Why the Answer Is Correct

Availability zones are physically separate datacenters within an Azure region, and enabling zones on an AKS cluster spreads the worker nodes (and their node pools) across three independent zone locations. That directly satisfies the first two requirements: zonal redundancy keeps pods schedulable when one datacenter is offline for planned maintenance, and it keeps the application serving traffic when an entire zone fails unexpectedly. AKS also performs zone-balanced scheduling so that scale-out adds nodes evenly across zones, which pairs with the cluster autoscaler to satisfy the dynamic-scaling requirement without concentrating new capacity in a single datacenter. Node image and Kubernetes version upgrades are performed with surge logic against a zone-spread node set, so maintenance proceeds while remaining replicas stay healthy. This is the only option in the list that both keeps the cluster inside one region and removes the single-datacenter dependency.

Why the Other Options Are Wrong

Option A (a single availability zone) places every node in one physical datacenter, so planned maintenance in that facility and any datacenter failure both take the cluster down — the exact scenario the question asks you to design against. Option B adds multiple node pools, but as one community member pointed out, multiple pools "provide workload separation and can scale independently, but they do not provide datacenter-level resiliency when they all reside in the same zone"; this is the most tempting distractor because the word "multiple" implies redundancy that node pools do not actually deliver. Option C goes too far in the wrong direction: an AKS cluster is region-scoped and its nodes cannot be deployed to different Azure regions, so multi-region would require multiple clusters plus global traffic routing, adding complexity that the question's single-cluster requirement does not ask for. Only D addresses the datacenter-failure and maintenance requirements at the zone level the question describes.

Community Comment Notes

Ras1364 answered D and argued explicitly against B, noting that pools in one zone cannot deliver datacenter-level resiliency and that Microsoft documents availability zones as physically distributing AKS resources across separate zones. gpais also selected D, explaining that enabling availability zones "spreads the AKS cluster's worker nodes across multiple physically separated datacenters within a single region" so that if one zone fails the remaining nodes stay healthy and downtime is minimized. Both comments converge on the same reasoning the requirements point to, and neither surfaces a credible argument for the single-zone or cross-region options. Use this consensus as confirmation, but the decisive evidence remains the requirements themselves: maintenance tolerance, dynamic scaling, and datacenter-failure tolerance all map to zone redundancy.