AKS Node / Pod / Cluster Sizing Calculator

Estimate replica counts, node pool sizes, and cluster tier for a two-service workload (frontend/BFF + API) on AKS from load-test inputs. Name each service and pick its runtime — sizing guidance adapts. Adjust any field — results update live.

1 · Service workloads

UI + BFF

Handles UI rendering and backend-for-frontend logic in one service.

Service name

Runtime / stack

Peak requests / sec (expected)

RPS one pod handles at target CPU

Minimum replicas (HA floor)

Pod CPU request (millicores)

Pod memory request (Mi)

Scale-out headroom multiplier

Buffer above raw RPS math so HPA scales before saturation.

API service

Backend API service, deployed separately from the other service.

Service name

Runtime / stack

Peak requests / sec (expected)

RPS one pod handles at target CPU

Minimum replicas (HA floor)

Pod CPU request (millicores)

Pod memory request (Mi)

Scale-out headroom multiplier

2 · Cluster & node pools

UI+BFF node pool

VM size

API node pool

VM size

System node pool

Node count

VM size

Dedicated to kube-system (CoreDNS, metrics-server, CNI) — app pods are not scheduled here.

Shared node-pool assumptions

Reserved for OS/kubelet (%)

Azure kube-reserved + system-reserved + eviction threshold.

DaemonSet overhead / node (millicores)

CNI, monitoring agent, log shipper, etc.

DaemonSet overhead / node (Mi)

Extra headroom nodes per pool

So a node drain / upgrade doesn't cause a scheduling failure.

Availability zones

Need Kubernetes Long Term Support (compliance)

3 · Results

Node distribution across pools

Total nodes provisioned, by pool

Config summary

This is a planning estimate, not a guarantee — it only knows what you tell it. The load numbers (RPS/pod, CPU, memory) should come from real load tests against your service's actual memory/GC behavior for its runtime, not guesses. Re-run this after every load test, and validate the resulting HPA/node-pool config in a staging cluster before rolling to production.

UI + BFF

API service

UI+BFF node pool

API node pool

System node pool

Shared node-pool assumptions

Node distribution across pools

Config summary Copy

Config summary

AKS Sizing Calculator