SC-300 domain - 25% of the exam

Plan and Implement Workload Identities

Plan and Implement Workload Identities is 25% of the SC-300 exam. These are the objectives it covers, each with practice questions, with every answer explained.

The domain in numbers

70
Practice questions
6
Objectives
59
Medium
11
Hard

Where people go wrong

  • Giving a managed identity the wrong lifecycle

    The wrong answers pair a managed identity with the wrong lifecycle. A user-assigned managed identity is a standalone object that survives deletion of the virtual machine, so it leaves an orphan, while a system-assigned identity is deleted automatically with its host resource. Sharing runs the other way: a system-assigned identity is bound to one resource and cannot cover a fleet, whereas one user-assigned identity can be granted roles before the compute exists and attached to a virtual machine, an Azure Function and a Logic App alike.

    See it in a sample question
  • Offering delegated permissions to a daemon with no user

    The wrong answers offer delegated permissions or user consent to a background service. Delegated permissions act on behalf of a signed-in user, so a daemon with no user cannot use them. An application permission such as User.Read.All gives app-only access under the app's own identity, and it needs administrator consent. The same split applies to app roles: assigning a role to a user does not help a daemon, whereas assigning it to the daemon's service principal puts the role claim in its token.

    See it in a sample question
  • Using a Defender for Cloud Apps policy at the wrong moment

    The wrong answers pick the policy type that acts at the wrong moment. An access policy decides at sign-in whether a session may start, so it can block unregistered or unjoined devices outright but cannot block a single download. A session policy proxies the live session, so it can block a download, or scan uploads for malware, while the user keeps working. Activity policies, file policies and app connectors work on logged events or stored data after the fact.

    See it in a sample question
  • Swapping the consent mechanisms

    The wrong answers swap the consent mechanisms. The admin consent workflow lets blocked users request approval and routes each request to designated reviewers, so it fits a tenant that has disabled user consent. It does not enable self-consent. When a delegated permission must be authorised for all users in a single action, the step is granting tenant-wide administrator consent directly. Self-consent for selected permissions covers only delegated permissions classified as low impact, and broader user consent removes the central review.

    See it in a sample question
  • Using directory roles or groups where app roles belong

    The wrong answers borrow a role or grouping from the wrong place. Directory roles and administrative units govern administration of the tenant, and groups are general directory objects, so none is the purpose-built way to send an in-app authorisation value. App roles are defined on the app registration, assigned on the enterprise application under Users and groups, and returned as a roles claim. The App roles page only declares them and assigns nobody.

    See it in a sample question

Objectives in this domain

What to study

The official documents this domain's practice questions cite most.

Sample questions from this domain

Free samplePlan and Implement Workload Identitiesmedium

A virtual machine hosts one application whose entire purpose is to read messages from a single Azure Storage queue, and the identity team wants the workload's identity to be removed automatically when that virtual machine is deleted, leaving no orphaned identity object to clean up. Which managed identity configuration meets this lifecycle requirement?

  • AA user-assigned managed identity attached to the virtual machine, retained so it can be reused by future resources.
  • BA service principal created manually in Microsoft Entra ID and assigned to the virtual machine for the duration of its life.
  • CA system-assigned managed identity enabled on the virtual machine, which Azure deletes automatically when the virtual machine is deleted. Correct
  • DAn app registration with a client secret deployed to the virtual machine and deleted by a teardown script when the VM is removed.
A system-assigned managed identity is tied to its host resource's lifecycle and is deleted automatically when that resource is deleted. System-assigned managed identities are created and destroyed together with the single resource they belong to. Because the identity's lifecycle is bound to the virtual machine, deleting the VM removes the identity automatically, which is exactly why it suits a single-purpose workload that should leave no orphaned object behind.

Why A is wrong: A user-assigned identity is tempting because it is also credential-free, but its lifecycle is independent of the virtual machine, so it survives deletion and becomes the orphaned object the team wants to avoid.

Why B is wrong: A manually created service principal is plausible, but it has no automatic lifecycle binding to the virtual machine, so deleting the VM leaves the service principal behind to be cleaned up by hand.

Why C is correct: A system-assigned managed identity is tied directly to the lifecycle of its host resource, so Azure deletes the identity automatically when the virtual machine is deleted, leaving no orphaned object.

Why D is wrong: An app registration with a teardown script could be made to clean up, but it relies on a stored secret and a fragile manual script rather than the platform deleting the identity automatically.

Free samplePlan and Implement Workload Identitiesmedium

A background service runs with no signed-in user and must read all users' profiles from Microsoft Graph on a recurring schedule using its own identity. In the app registration's API permissions, which type of Microsoft Graph permission should the administrator configure so the service can call Graph under its own identity?

  • AAn application permission such as User.Read.All, consented by an administrator, which lets the app authenticate with the client credentials flow and read profiles under its own identity. Correct
  • BA delegated permission such as User.Read, granted with admin consent so the service inherits the rights of any user who has signed in to it recently.
  • CAn application permission such as User.Read.All that a tenant member self-approves through user consent, so the service can read all profiles without waiting for an administrator.
  • DA delegated permission such as User.ReadBasic.All with user consent enabled, since the service can prompt the scheduled task account to approve access at runtime.
Unattended app-only access to Microsoft Graph requires application permissions consented by an administrator and used through the client credentials flow. Microsoft Graph permissions come in two kinds. Delegated permissions let an app act on behalf of a signed-in user and need that user present in the flow. Application permissions grant the app its own identity to call Graph with no user, which is exactly what a background or daemon service uses through the client credentials flow. Application permissions are high privilege and always require an administrator to consent on behalf of the organisation. A scheduled service with no interactive user therefore needs an application permission such as User.Read.All with admin consent, not a delegated scope that depends on a user sign-in.

Why A is correct: Application permissions grant the app its own app-only access through the client credentials flow, require admin consent, and let an unattended service read all profiles without any signed-in user, matching the requirement.

Why B is wrong: Delegated permissions act on behalf of a signed-in user and require a user present in the flow, so they cannot work for an unattended background service that has no interactive user.

Why C is wrong: Application permissions are the right type for app-only access, but they always require admin consent and can never be self-approved through user consent, so this option configures the credential incorrectly.

Why D is wrong: User consent and delegated scopes still need an interactive user to authenticate and consent, which an unattended scheduled service does not have, so this flow cannot complete.

Free samplePlan and Implement Workload Identitieshard

A security team has routed a sanctioned SaaS application through Conditional Access app control. They want every file a user uploads into the application during the session to be inspected for malware in real time, and any infected file blocked before it reaches the app. Which Microsoft Defender for Cloud Apps configuration meets this requirement?

  • AA session policy set to control file upload, with the built-in malware detection inspection method applied to block matching files Correct
  • BAn access policy that blocks the user's sign-in whenever the upstream antivirus engine flags the device as infected before the session opens
  • CA Cloud Discovery anomaly detection policy that scores the app's traffic and quarantines uploads once the application's risk score crosses a threshold
  • DAn app connector via the API that scans files already stored in the application and quarantines any malware found after upload completes
A Defender for Cloud Apps session policy controlling file upload with malware inspection blocks infected files in real time inside a proxied session. Session policies place Defender for Cloud Apps inline as a reverse proxy over the live session, so the file upload activity can be evaluated and malware-detection inspection can block an infected file before it lands in the app, which neither an access policy at sign-in nor an after-the-fact API scan can do.

Why A is correct: A session policy proxies the live session, and the file upload activity type with malware inspection scans each uploaded file in real time and blocks infected ones.

Why B is wrong: An access policy decides whether a session may start; it operates at sign-in and cannot inspect individual files being uploaded inside the live session.

Why C is wrong: Cloud Discovery scores shadow IT from traffic logs and never sits inline on the session, so it cannot inspect or block an individual upload as it happens.

Why D is wrong: API app connectors scan data at rest after the fact, which is tempting but acts post-upload rather than blocking the infected file in real time during the session.

Free samplePlan and Implement Workload Identitiesmedium

After the identity team disabled user consent in Microsoft Entra ID, employees who need a new SaaS application can no longer grant it permissions and are asking IT to approve apps. The team wants users to submit a justified request that designated reviewers can approve, with the approved consent granted tenant-wide. Which feature should the administrator enable?

  • AAn entitlement management access package containing the SaaS application, which users request and an approver grants so the app permissions are consented.
  • BThe admin consent workflow, which lets users request admin consent for an application and routes the request to nominated reviewers who approve or deny it. Correct
  • CA Conditional Access policy with an approval grant control, which prompts a reviewer to approve before the user is allowed to consent to the application.
  • DThe user consent settings set to verified publishers, which lets users self-consent so that requests no longer need to reach an administrator at all.
The admin consent workflow lets users request consent for apps and routes those requests to designated reviewers who grant tenant-wide consent. When user consent is disabled, the admin consent workflow restores a path forward: a blocked user sees a Request approval option, supplies a business justification, and the request is sent to the reviewers the administrator nominated. An approving reviewer grants admin consent for the whole tenant. Access packages govern resource assignment not consent, Conditional Access lacks a consent-approval control, and re-enabling self-consent removes the review step.

Why A is wrong: Access packages grant assignment and access to resources but do not perform application permission consent, so they cannot replace the admin consent decision that the blocked apps require.

Why B is correct: The admin consent workflow gives blocked users a Request approval button, captures a justification, and routes it to designated reviewers whose approval grants tenant-wide admin consent, matching the stated need exactly.

Why C is wrong: Conditional Access governs sign-in conditions and has no permission-consent approval control, so it cannot capture a consent request or grant tenant-wide consent on a reviewer's approval.

Why D is wrong: Re-enabling self-consent contradicts the decision to disable user consent and removes the reviewer approval step the team wants, so it does not provide the request-and-approve flow described.

Free samplePlan and Implement Workload Identitiesmedium

A line-of-business application defines three app roles in its manifest: Reader, Contributor, and Administrator. The identity team must grant the Sales team Contributor-level access and the Operations team Administrator-level access through Microsoft Entra ID, so the application receives the correct role claim per group. How should the administrator assign these roles?

  • AEdit the application registration manifest and add the Sales and Operations group object IDs directly into the appRoles array so each role maps to a group.
  • BCreate two Microsoft Entra directory roles named Contributor and Administrator and assign the Sales and Operations groups to those directory roles.
  • CAdd the Sales and Operations groups as owners of the enterprise application so that ownership maps each group to its corresponding Contributor or Administrator role.
  • DUnder the enterprise application's Users and groups, add the Sales group with the Contributor app role and the Operations group with the Administrator app role. Correct
App roles defined in an application's manifest are assigned to users and groups on the enterprise application, which then issues the matching role claim. An application declares app roles in its registration manifest, but the assignment of those roles to principals happens on the enterprise application under Users and groups. When a member of an assigned group signs in, Microsoft Entra ID emits the roles claim for that app role. Manifest entries only define roles, directory roles administer the tenant, and ownership manages app configuration rather than issuing role claims.

Why A is wrong: The appRoles array in the manifest only defines the roles that exist; it does not assign them to principals, and putting group IDs there is invalid, so no role claim would be issued.

Why B is wrong: Microsoft Entra directory roles govern administration of the tenant, not claims sent to a specific application, so they cannot deliver an app-defined Contributor or Administrator role claim.

Why C is wrong: Application ownership grants rights to manage the app's configuration; it does not assign app roles or send role claims, so ownership cannot convey Contributor or Administrator access to the application.

Why D is correct: App role assignment is performed on the enterprise application under Users and groups, where each user or group is given a specific app role, so the application receives the matching role claim for each group at sign-in.

Other domains in this exam

See also the SC-300 cert hub, the study guide, and the cheat sheet.

Examworthy is not affiliated with or endorsed by Microsoft. Original, blueprint-aligned practice material only.