Azure IV: Governance, Storage & Backup
Running an Azure estate for the AZ-104 and for real: Entra users, licences, guests and SSPR; management groups, Azure Policy with deny, modify and remediation, resource locks and the cost tools; storage accounts in depth (redundancy, tiers, keys, SAS, firewalls, encryption with your own key, versioning, soft delete, lifecycle, object replication, Azure Files, AzCopy); Azure Backup and Site Recovery - with the incidents each of them causes.
In plain words
Think of a big office building. Someone has to hand out badges and take them back (Entra ID), decide which floors belong to which department (management groups and resource groups), put up house rules that the doors enforce on their own (Azure Policy), bolt down the server racks (locks), and watch the electricity bill (budgets and Advisor).
The building also has a warehouse (storage accounts) with keys, guest passes, alarms and a shredder that runs on a timer, and an insurance policy with a second building across town (Backup and Site Recovery). This chapter is the facilities team's handbook.
Why it matters on call
This is the half of Azure administration that never shows up in an architecture diagram and causes a large share of the incidents: the policy that blocks a release, the lock that stops a cleanup, the SAS token pasted into a ticket, the lifecycle rule that archives a file the app still reads, the backup that silently stopped working nine days ago.
It is also most of the AZ-104: identities and governance, storage, and backup and recovery are roughly half of the exam. Every lab here is something you will do, or wish someone had done, on a real platform team.
Lessons
- 32.1Microsoft Entra ID: users, groups, licences, guests and SSPR
- 32.4Management groups, subscriptions, resource groups and tags at scale
- 32.7Azure Policy: rules that say no (deny, audit, scopes and exclusions)
- 32.12Policy that fixes things: modify, deployIfNotExists, remediation, initiatives and your own rules
- 32.15Resource locks: the last guard before delete
- 32.19Costs: budgets, cost alerts and Azure Advisor
- 32.21Storage accounts in depth: performance, redundancy, tiers and encryption
- 32.24Access to storage: keys, SAS, stored access policies, data roles and firewalls
- 32.29Protecting blob data: soft delete, versioning, lifecycle and object replication
- 32.34Azure Files, identity-based access, and moving data with AzCopy
- 32.36Azure Backup: vaults, policies, backup and restore
- 32.41Site Recovery: replicate, test, fail over, re-protect
- 32.44Watching backups: alerts, reports, and a vault attackers can't empty
Open this chapter in the terminal Free, in your browser - a real Ubuntu terminal to try it in, with missions that check your work.
Questions people ask
Do I need Azure I-III before this chapter?
They help: Azure I taught the CLI, RBAC and storage basics, Azure II networking and private endpoints, Azure III Azure Monitor and KQL, and all of those come back here. The chapter works in the same simulated tenant, so the vault, the workspace and the network you met there are the ones you use now.
Is this enough for the AZ-104?
It covers the exam's identity and governance domain, the storage domain and the backup and recovery part of monitoring. Compute, networking and Azure Monitor come in Azure V together with the exam lesson and mock exams. The question bank on the Certifications page maps every objective to the lessons and labs that teach it.
Why are some things done in the portal?
Because that is the only way Azure offers them, or by far the normal one: self-service password reset settings and Site Recovery for Azure VMs have no general-availability CLI. The Console's Cloud portal pages do those steps and show the command they ran (azportal, labelled simulator), so the lab still records what you did.
How realistic is the simulated Azure?
Command names, options, output shapes and error texts follow azure-cli 2.91 and Microsoft Learn: RequestDisallowedByPolicy with its evaluation details, ScopeLocked, the storage 403 codes, BlobArchived, the backup job errors. The tenant behind them is simulated, with time compressed where Azure is slow (compliance scans, rehydration, replication). Anything the lab had to invent is marked (simulator).
Which part matters most on call?
Reading errors and knowing which control refused something: a lock, a policy, RBAC, the storage firewall, an expired or wrongly signed SAS. The second is backups: knowing that alerts are routed, that the last good recovery point is recent, and how long a restore really takes. The incidents in this chapter are built around exactly those moments.