Opscraft
Azure for AWS engineers · Lesson 1 of 9 · 3 min read

Azure fundamentals and the resource hierarchy

Azure appears as "good to have" in many Bangalore DevOps postings, especially at GCCs of European and US enterprises that run Microsoft-heavy estates. You don't need production depth to pass — you need to map Azure concepts onto your AWS knowledge accurately, speak the vocabulary correctly (resource groups, Entra ID, managed identities, VNets, NSGs, AKS), and show you could be productive quickly. This optional topic is a complete orientation written for an AWS engineer, with the AKS and Terraform details interviewers ask about.

The hierarchy

Enlarge diagramEntra tenantthe organisation's identity directoryRoot management groupManagement groupsPlatform, Prod, Non-prod, Sandbox (≈ AWS OUs)Subscriptionsbilling, quota, access boundary (≈ AWS account)Resource groupsresources that live and die togetherResourcesVMs, VNets, AKS clusters, storage accounts
Figure 1. The Azure resource hierarchy. Policy and access assigned higher up are inherited by everything below.
Swipe sideways to see all of it, or tap to open it full screen.
  • Tenant — an instance of Microsoft Entra ID (formerly Azure Active Directory) representing an organisation; all identities live here.
  • Management groups — organise subscriptions; Azure Policy and RBAC assigned here are inherited downwards. (≈ AWS Organizations OUs.)
  • Subscription — the billing and quota unit and an access boundary (≈ an AWS account).
  • Resource group (RG) — a logical container for resources that share a lifecycle and permissions: deploy, manage and delete them together (az group delete removes everything inside). Every resource lives in exactly one RG. AWS has no direct equivalent — the closest are tags or CloudFormation stacks.
  • Resource — an individual service instance, identified by a resource ID: /subscriptions/<id>/resourceGroups/rg-shop-prod/providers/Microsoft.ContainerService/managedClusters/aks-shop-prod.

All management goes through Azure Resource Manager (ARM) — the control plane API used by the portal, CLI, PowerShell, ARM/Bicep templates and Terraform.

Regions and availability

  • Regions (India has Central India — Pune, South India — Chennai, West India — Mumbai) contain availability zones — physically separate datacentres with independent power, cooling and networking (not every region has zones).
  • Region pairs — most regions are paired with another in the same geography (Central India ↔ South India) for disaster recovery: platform updates are staggered across pairs, and geo-redundant storage replicates to the pair.
  • Services are zonal (pinned to a zone), zone-redundant (spread automatically across zones), or regional.
Enlarge diagramCENTRAL INDIA (PUNE)SOUTH INDIA (CHENNAI)zone 1zone 2zone 3paired regiongeo-redundant copies, staggered updates
Figure 2. Availability in Azure: zones inside a region for local failures, a paired region for disasters.
Swipe sideways to see all of it, or tap to open it full screen.