AZ-305 Practice Question 58: AKS Pod IP Exhaustion
Microsoft Azure Solutions Architect Expert (AZ-305) · Test 6 · question 8 of 10
An AKS cluster will grow to about 4,000 nodes running roughly 100,000 pods. The virtual network address space is limited because it connects to on-premises networks over ExpressRoute, and the network team refuses to reserve a large VNet range for pod addresses. The platform team does not want to create and maintain route tables for pod traffic, and it wants a networking model that AKS will continue to support.
Which networking option should you choose?
Show answer and explanation
Answer:
With Azure CNI Overlay, only the nodes receive addresses from the VNet subnet. Pods take addresses from a separate private CIDR that is logically different from the virtual network, which avoids the address exhaustion of the flat models and needs no route tables to be managed. It also scales further than kubenet, which is the reason it is the recommended replacement.
Option A assigns a VNet address to every pod. For 100,000 pods that would consume a very large range, which is the constraint the network team refuses.
Option B does use a separate pod address space, but it needs user-defined routes, has lower scale limits, and AKS is ending support for kubenet on 31 March 2028, so it fails the requirement about continued support.
Option D still takes pod addresses from VNet subnets, only from a dedicated one, so the address consumption problem remains.