Azure Availability Zone Planner

VMs per zone = ceil(required ÷ (zones − tolerated losses)), total = per-zone × zones; composite availability = 1 − (1 − single-zone availability)^(tolerated losses + 1). Zone-redundant sizing works backwards from the failure you tolerate: if one of three zones can go, each surviving pair must carry the whole peak, so each zone holds half the load and you deploy 1.5× the capacity. Because zones are independent failure domains, spreading across them multiplies the failure probabilities, which is how the published target rises from 99.9% for a single VM to 99.99% across zones.

The overhead figure is the real price of a 99.99% target, and it is paid every hour whether a zone fails or not; cross-zone traffic is metered too, so chatty services that gossip between zones add a data charge on top of the compute overhead.

This calculator takes 7 inputs: VMs needed to carry peak load, Zones used, Zones that may fail simultaneously, Availability of one zone deployment, VM rate, Cross-zone traffic, Cross-zone data price per GB. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.