Fault Domain
A logical grouping representing shared physical infrastructure (power source and network switch) within a datacenter; VMs in different fault domains within an Availability Set won't fail together from a single hardware fault.
Fault Domain
A fault domain maps to a physical rack sharing power and networking. Spreading VMs across fault domains means one rack-level failure doesn't take out your whole application tier.
Why It Matters in Production
Within Razorpay's payments Availability Set, VMs are automatically spread across 3 fault domains, so a single PDU failure only affects roughly a third of instances.
az vm availability-set show --name payments-avset \ --resource-group razorpay-prod-rg --query platformFaultDomainCountRememberAzure assigns fault domain placement automatically within an Availability Set — you cannot manually pin a specific VM to a specific fault domain.
Frequently Asked Questions
How many fault domains does an Azure Availability Set typically span?
An Availability Set spreads VMs across up to 3 fault domains by default (configurable up to a region's maximum, often 3), each representing a separate physical rack with independent power and network switch. This means a single power supply or top-of-rack switch failure can only take down VMs in one fault domain, not the whole set, provided your VMs are actually distributed across more than one.
What's the difference between using Fault Domains via an Availability Set versus Availability Zones?
Fault Domains inside an Availability Set protect against rack-level hardware failure within a single datacenter, while Availability Zones protect against failure of an entire datacenter (power, cooling, networking at the building level) by spreading VMs across physically separate facilities within a region. For higher resilience requirements, Availability Zones are the stronger option; Availability Sets with fault domains are the lower-cost, same-datacenter alternative.