Sign in
Home/Microsoft/Microsoft Azure Administrator (AZ-104)/Free questions

Microsoft Azure Administrator (AZ-104) — Free Practice Questions

10 free sample questions from a bank of 668, with the correct answers and explanations. No signup required — start practising right now.

1Your company has serval departments. Each department has a number of virtual machines (VMs). The company has an Azure subscription that contains a resource group named RG1. All VMs are located in RG1. You want to associate each VM with its respective department. What should you do?
  • Create Azure Management Groups for each department.
  • Create a resource group for each department.
  • Assign tags to the virtual machines.
  • Modify the settings of the virtual machines.
Answer: C

The short version

C — Assign tags to link each VM to its department. Tags add department metadata without moving resources, enabling filtering, chargeback and policy scoping inside RG1.

Key concepts in this question

  • Azure tags: Name-value pairs on resources and groups for organization, cost tracking and automation targeting.
  • Single resource group constraint: All VMs stay in RG1, so department separation must be logical rather than physical.
  • Management groups contrast: Hierarchy for governance across subscriptions, overkill for labeling VMs in one group.
  • Resource group split contrast: Moving VMs to per-department groups would satisfy grouping but violates the stated RG1 retention and adds move overhead.

Why C is correct

Tags associate each VM with its department while preserving RG1 placement, supporting views, cost allocation and targeted automation. They impose no relocation or hierarchy change. That metadata-without-move fit makes tagging the settled action.

Why the others are wrong

  • A. Management Groups are wrong because they govern subscriptions and policy inheritance, not per-VM department labeling.
  • B. New resource groups are wrong because they require moves and contradict keeping all VMs in RG1.
  • D. Modify VM settings is wrong because size or config tweaks carry no department association semantics.

AZ-104 exam tip

When the stem says associate or organize without moving, choose tags, reserving resource groups for lifecycle boundaries and management groups for multi-subscription governance.

2Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company has an Azure Active Directory (Azure AD) subscription. You want to implement an Azure AD conditional access policy. The policy must be configured to require members of the Global Administrators group to use Multi-Factor Authentication and an Azure AD-joined device when they connect to Azure AD from untrusted locations. Solution: You access the multi-factor authentication page to alter the user settings. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, MFA user settings cannot enforce device plus location policy. Requiring MFA and hybrid-joined devices from untrusted locations needs Conditional Access grant controls, not per-user MFA toggles.

Key concepts in this question

  • Conditional Access: Policy engine combining users, locations, devices and risk with grant requirements such as MFA and device compliance.
  • Per-user MFA settings: Legacy enforcement for MFA only, unaware of device state or network location.
  • Grant versus session: Grant controls decide access requirements, while session controls limit experience after access.
  • Untrusted location plus device rule: Demands both signals evaluated together, only possible in Conditional Access.

Why B is correct

Altering MFA user settings enables two-step verification but cannot test for Azure AD-joined devices or untrusted networks. The goal explicitly requires both conditions with MFA, which lives under grant controls. Because the solution uses the wrong blade, it does not meet the goal as settled.

Why the others are wrong

  • A. Yes is wrong because per-user MFA lacks device and location evaluation, so Global Admins could sign in without a joined device.

AZ-104 exam tip

Map MFA page to MFA-only, session to experience limits, and grant to who plus what-device plus where, picking grant for combined MFA-device-location stems.

3Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company has an Azure Active Directory (Azure AD) subscription. You want to implement an Azure AD conditional access policy. The policy must be configured to require members of the Global Administrators group to use Multi-Factor Authentication and an Azure AD-joined device when they connect to Azure AD from untrusted locations. Solution: You access the Azure portal to alter the session control of the Azure AD conditional access policy. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, session controls cannot require MFA and joined devices. Session settings manage sign-in frequency and persistence, while access requirements belong to grant controls.

Key concepts in this question

  • Grant controls: Enforce MFA, device compliance, hybrid join or approved apps before granting access.
  • Session controls: Constrain the session after grant, such as sign-in frequency, persistent browsing and app-enforced restrictions.
  • Global Admin protection: Needs strong grant policy scoped to role, location and device posture.
  • Portal blade distinction: Choosing the wrong control category leaves the stated device and MFA demand unenforced.

Why B is correct

Changing session controls adjusts lifetime and experience but never imposes MFA or device-join prerequisites. The requirement calls for both signals at access time, which only grant controls provide. Since the solution edits the unrelated category, it fails the goal as settled.

Why the others are wrong

  • A. Yes is wrong because session tweaks leave Global Admins able to connect without MFA or a joined device from untrusted places.

AZ-104 exam tip

Remember grant equals get-in requirements and session equals stay-in limits, so MFA-plus-device stems always need grant, never session.

4Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company has an Azure Active Directory (Azure AD) subscription. You want to implement an Azure AD conditional access policy. The policy must be configured to require members of the Global Administrators group to use Multi-Factor Authentication and an Azure AD-joined device when they connect to Azure AD from untrusted locations. Solution: You access the Azure portal to alter the grant control of the Azure AD conditional access policy. Does the solution meet the goal?
  • Yes
  • No
Answer: A

The short version

A — Yes, grant controls can enforce MFA plus joined device. Editing grant controls lets the policy demand both signals for Global Admins from untrusted locations.

Key concepts in this question

  • Grant control options: Require MFA, require compliant or hybrid-joined device, require approved app, applied with AND or OR logic.
  • Targeting: Scoped to Global Admin role, untrusted locations, and cloud app entry for precise coverage.
  • Device trust: Hybrid-joined status proves corporate ownership and management before cloud access.
  • Correct blade fit: Portal grant section is the documented place for combined authentication and device rules.

Why A is correct

Altering grant controls directly implements require MFA and require hybrid-joined device under the stated location condition. Evaluation occurs before token issuance, blocking noncompliant attempts. That exact capability match makes Yes the settled evaluation.

Why the others are wrong

  • B. No is wrong because grant controls do support joint MFA and device enforcement, fully meeting the described goal.

Grant controls evaluate role, location and device signals together before issuing tokens, which is precisely the combined enforcement the scenario demands for privileged access.

AZ-104 exam tip

Spot grant in the solution for MFA-plus-device-location scenarios and approve it, since grant is the only control family that combines those demands.

5You are planning to deploy an Ubuntu Server virtual machine to your company's Azure subscription. You are required to implement a custom deployment that includes adding a particular trusted root certification authority (CA). Which of the following should you use to create the virtual machine?
  • The New-AzureRmVm cmdlet.
  • The New-AzVM cmdlet.
  • The Create-AzVM cmdlet.
  • The az vm create command.
Answer: D

The short version

D — Use az vm create for custom Ubuntu with trusted root CA. CLI creation supports cloud-init and custom data for certificates, while legacy or invented cmdlets lack that path.

Key concepts in this question

  • Custom deployment need: Injecting a trusted root CA requires provisioning-time scripts or cloud-init payloads.
  • AZ CLI flexibility: az vm create accepts custom data, extension settings and image plans suited to Ubuntu customization.
  • PowerShell cmdlet status: New-AzVM provisions VMs but the listed legacy and invented variants are outdated or nonexistent.
  • CA trust mechanics: Adding the authority to system stores ensures workloads validate internal services without errors.

Why D is correct

Only az vm create in the options aligns with a custom Linux deployment carrying extra CA configuration via supported parameters. Legacy AzureRm modules are retired and Create-AzVM is not the valid creation cmdlet. That deploy-plus-customize fit makes D the settled choice.

Why the others are wrong

  • A. New-AzureRmVm is wrong because the AzureRm module is deprecated and unsuitable for new custom deployments.
  • B. New-AzVM is wrong for this stem because the tested custom CA path favors CLI with cloud-init in the settled key.
  • C. Create-AzVM is wrong because it is not a valid creation cmdlet name in Az PowerShell.

AZ-104 exam tip

Link custom Linux plus cloud-init to az vm create, and treat AzureRm plus Create-AzVM as distractor naming traps.

6Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company makes use of Multi-Factor Authentication for when users are not in the office. The Per Authentication option has been configured as the usage model. After the acquisition of a smaller business and the addition of the new staff to Azure Active Directory (Azure AD) obtains a different company and adding the new employees to Azure Active Directory (Azure AD), you are informed that these employees should also make use of Multi-Factor Authentication. To achieve this, the Per Enabled User setting must be set for the usage model. Solution: You reconfigure the existing usage model via the Azure portal. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, you cannot reconfigure the MFA usage model in place. Per-Authentication versus Per-Enabled-User is fixed at creation for legacy MFA providers, requiring a new provider.

Key concepts in this question

  • Legacy MFA models: Per-Auth bills per challenge and Per-User licenses enabled users, chosen when the provider is created.
  • Immutability rule: Existing provider model cannot flip through portal toggles after deployment.
  • Migration path: Create a new provider with the desired model and migrate users and settings.
  • Portal edit trap: Reconfiguring the same object appears plausible but violates the model-lock behavior.

Why B is correct

The solution tries to reconfigure the existing usage model via the portal, which the platform blocks because the model is set at creation. Meeting the new Per-Enabled-User need demands a replacement provider. Since the proposed edit cannot effect the change, No is the settled evaluation.

Why the others are wrong

  • A. Yes is wrong because portal reconfiguration of the same provider does not convert billing models.

Documentation stresses planning a new provider and staged migration rather than editing the locked property, since billing and enablement flow from the original model choice.

AZ-104 exam tip

Treat usage-model change as new-provider work, so any reconfigure-existing-model solution earns No.

7Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company's Azure solution makes use of Multi-Factor Authentication for when users are not in the office. The Per Authentication option has been configured as the usage model. After the acquisition of a smaller business and the addition of the new staff to Azure Active Directory (Azure AD) obtains a different company and adding the new employees to Azure Active Directory (Azure AD), you are informed that these employees should also make use of Multi-Factor Authentication. To achieve this, the Per Enabled User setting must be set for the usage model. Solution: You reconfigure the existing usage model via the Azure CLI. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, Azure CLI cannot convert the existing MFA model. Like the portal, CLI edits cannot change a locked Per-Authentication provider into Per-Enabled-User.

Key concepts in this question

  • Model lock: Billing model binds to the MFA provider object at provisioning time.
  • Tooling equivalence: Portal and CLI share the same backend validation, so neither bypasses immutability.
  • Correct remediation: Provision a new provider with Per-Enabled-User and transition users.
  • CLI misconception: Command-line power does not override service constraints on provider conversion.

Why B is correct

Reconfiguring via CLI targets the same immutable property as the portal attempt, so it fails for the same reason. The new employees need coverage under a different model, which requires replacement rather than edit. That platform limit makes No the settled answer.

Why the others are wrong

  • A. Yes is wrong because CLI reconfiguration still hits the fixed-model validation and leaves billing unchanged.

Both interfaces call the same provider management API, so switching tools cannot unlock a property that the service marks immutable after creation.

AZ-104 exam tip

Remember tool does not trump model lock, so CLI plus reconfigure-existing still means No for usage-model changes.

8Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company's Azure solution makes use of Multi-Factor Authentication for when users are not in the office. The Per Authentication option has been configured as the usage model. After the acquisition of a smaller business and the addition of the new staff to Azure Active Directory (Azure AD) obtains a different company and adding the new employees to Azure Active Directory (Azure AD), you are informed that these employees should also make use of Multi-Factor Authentication. To achieve this, the Per Enabled User setting must be set for the usage model. Solution: You create a new Multi-Factor Authentication provider with a backup from the existing Multi-Factor Authentication provider data. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, new provider with backup from old data does not change the model need. Creating another provider without selecting the required Per-Enabled-User model leaves the goal unmet.

Key concepts in this question

  • Provider data backup: Copies configuration for continuity, but model choice still governs billing behavior.
  • Requirement specificity: New staff must use MFA under Per-Enabled-User, demanding that exact model.
  • Incomplete solution: Backup and recreate without model correction repeats the original mismatch.
  • Evaluation focus: Whether the proposal delivers the required usage model, not merely whether a new object exists.

Why B is correct

The solution creates a provider from backup data yet does not establish Per-Enabled-User as required. Without that model, the new employees remain improperly licensed for the intended scheme. Because the core requirement stays unsatisfied, No is the settled result.

Why the others are wrong

  • A. Yes is wrong because a cloned provider inheriting the old model cannot satisfy the new model mandate.

Evaluators must confirm the new object uses Per-Enabled-User licensing, otherwise the solution only duplicates infrastructure without fixing the business requirement.

AZ-104 exam tip

Check new-provider proposals for the required model name, since backup alone never implies the billing model changed.

9Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company has an Azure Active Directory (Azure AD) tenant named weyland.com that is configured for hybrid coexistence with the on-premises Active Directory domain. You have a server named DirSync1 that is configured as a DirSync server. You create a new user account in the on-premise Active Directory. You now need to replicate the user information to Azure AD immediately. Solution: You run the Start-ADSyncSyncCycle -PolicyType Initial PowerShell cmdlet. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, Initial sync is wrong for immediate single-user replication. Delta sync replicates recent on-premises changes quickly, while Initial performs a full import and is unnecessarily heavy.

Key concepts in this question

  • ADSync cycles: Delta handles incremental changes in minutes, Initial rebuilds the connector space fully.
  • Urgent user need: New account must appear in Entra ID promptly without full resynchronization.
  • Cmdlet choice: Start-ADSyncSyncCycle with Delta targets pending deltas, Initial triggers complete evaluation.
  • Overkill and risk: Full sync wastes time and can throttle, opposite of immediate replication.

Why B is correct

Running Initial when only one new user needs replication initiates a full cycle instead of the lightweight delta. Best practice for immediate needs is Delta, not Initial. Because the solution selects the heavyweight policy type, it does not meet the goal as settled.

Why the others are wrong

  • A. Yes is wrong because Initial delays targeted replication with full processing rather than accelerating the single object.

Delta cycles process pending imports and exports in minutes, making them the standard for urgent single-object replication without full re-evaluation.

AZ-104 exam tip

Pair immediate single-object sync with Delta and full rebuild with Initial, so Initial for urgent deltas always means No.

10Note: The question is included in a number of questions that depicts the identical set-up. However, every question has a distinctive result. Establish if the solution satisfies the requirements. Your company has an Azure Active Directory (Azure AD) tenant named weyland.com that is configured for hybrid coexistence with the on-premises Active Directory domain. You have a server named DirSync1 that is configured as a DirSync server. You create a new user account in the on-premise Active Directory. You now need to replicate the user information to Azure AD immediately. Solution: You use Active Directory Sites and Services to force replication of the Global Catalog on a domain controller. Does the solution meet the goal?
  • Yes
  • No
Answer: B

The short version

B — No, Sites and Services cannot push to Entra ID. Forcing Global Catalog replication stays on-premises and never triggers Azure AD Connect cloud export.

Key concepts in this question

  • On-prem replication: Sites and Services moves directory data between domain controllers only.
  • Cloud sync boundary: Only Azure AD Connect with a sync cycle exports to the Entra tenant.
  • DirSync server role: Holds the scheduler and connectors that translate on-prem changes for cloud provisioning.
  • Immediate need: Requires Delta sync execution, not intra-site topology forcing.

Why B is correct

Global Catalog convergence does not invoke the sync engine export to weyland tenant, so the new user stays absent from Entra ID. The correct action is a Delta cycle on DirSync1. Since the proposal acts on the wrong tier, it fails the goal as settled.

Why the others are wrong

  • A. Yes is wrong because on-prem convergence alone creates no cloud object without a Connect export cycle.

Only a Connect export creates or updates the cloud identity, so on-premises convergence must still be followed by a Delta cycle on the sync server.

AZ-104 exam tip

Separate on-prem replication from cloud sync, choosing ADSync Delta for to-Entra urgency and Sites and Services only for DC-to-DC convergence.

Want the full bank of 668 questions for Microsoft Azure Administrator (AZ-104)? See all practice exams.