Q1.An Azure subscription contains the following resources.
– An Azure SQL Database logical server named Server1, which hosts a database named DB1.
– An Azure SQL Managed Instance named Instance1, which hosts a database named DB2.
You plan to configure database auditing for these resources.
The solution must meet the following requirements.
– Allow the audit data to be reviewed centrally in a location where KQL queries can be run.
– Minimize the ongoing operational overhead when new databases are added.
What should you configure?
Select the appropriate options in the answer area.
Show answer
When auditing is enabled at the server or instance level, the audit settings are applied collectively to all databases beneath it, so there is no need to repeat individual configuration as databases are added later, which keeps the management burden low.
In addition, when the audit logs are sent to a Log Analytics workspace, audit data from multiple resources can be stored centrally and searched and analyzed with KQL in Azure Monitor Logs.
Event Hubs is well suited to streaming to external systems, but it is not a destination for analyzing directly with KQL.
A Storage account is well suited to long-term retention, but Log Analytics is the best fit for the requirement of centralized analysis with KQL.
Auditing policy at the server and database levels
Set up auditing for Azure SQL Database and Azure Synapse Analytics
Q2.A company is considering a process to verify the safety of content generated by generative AI before displaying it to users.
The goal is to automatically detect harmful, inappropriate, or policy-violating responses.
Which capability should be adopted as a priority?
Show answer
Content filtering is a capability that analyzes the input to the AI and the generated output to detect or block harmful content such as violence, sexual content, self-harm, and hate.
In addition, safety evaluation lets you continuously measure the degree of harmfulness and policy violations in the responses of a generative AI application.
Increasing the context window relates to the amount of information that can be processed, adding GPUs relates to performance, and virtual networks relate to isolating and protecting communications, but none of these directly assess the safety of the generated content itself.
With Microsoft, you can use Azure AI Content Safety to detect harmfulness in user-generated and AI-generated content.
What is Azure AI Content Safety?
Risk and safety evaluators
Q3.A company is concerned about prompt injection attacks targeting an AI chatbot that integrates with core business systems.
Which is the most effective mitigation?
Show answer
Prompt injection is an attack that uses malicious input to manipulate the model’s instructions and behavior, causing it to misuse connected data and tools.
Therefore, it is important not to trust external input and to detect malicious instructions with validation and capabilities such as Prompt Shields.
Furthermore, in case an attack slips past detection, limiting the tools the AI can call and their execution permissions to the minimum necessary reduces the impact.
Disabling logging reduces investigative capability, and trusting responses unconditionally is dangerous.
The temperature parameter adjusts the diversity of the output and is not a countermeasure against attacks.
Prompt Shields
Artificial intelligence security
Q4.You are configuring a new Microsoft Sentinel workspace named Workspace1.
Your organization has an external IT service management (ITSM) system that is not supported by the Microsoft Sentinel solutions in Azure Marketplace.
Whenever a new security incident occurs in Workspace1, a service ticket must be automatically created in that ITSM system.
What should you create?
Show answer
A Microsoft Sentinel playbook is an automated workflow built on Azure Logic Apps.
Triggered by a new incident, it can call the API of an external ITSM system or an available Logic Apps connector to automatically create a service ticket.
Even if there is no packaged Microsoft Sentinel solution for the target ITSM product, integration is possible as a custom workflow.
In practice, you also configure an automation rule that runs the playbook when an incident is created.
A workbook is for visualization, a watchlist is for managing reference data, and an analytics rule is a capability for detecting threats and generating alerts and incidents, so they are not suited to creating tickets in an external system.
Automate threat response with playbooks in Microsoft Sentinel
Azure Logic Apps for Microsoft Sentinel playbooks
Q5.You have three internet-facing Azure App Service web apps named App1, App2, and App3.
All of the apps use built-in authentication.
Among them, App2 hosts a backend API.
Some internal users who should not be able to use that API are able to sign in to App2.
You must restrict access to App2 to only assigned Microsoft Entra users and groups.
What should you configure for App2?
To answer, drag the appropriate configuration to the correct method.
Each configuration may be used once, more than once, or not at all.
You might need to drag the split bar between panes or scroll to view content.
Drag each option to the matching requirement’s drop zone, or tap to place it. Click a filled zone to send it back. (The same option may be used more than once.)
IP access restrictions
Local authentication
Microsoft identity provider
Role-based access control (RBAC) assignment
Tenant-wide admin consent
Show answer
For App2, configure the Microsoft identity provider as the identity provider for built-in authentication, and authenticate users with Microsoft Entra ID.
In addition, when you set “Assignment required?” to Yes on the corresponding enterprise application, only explicitly assigned users or groups can sign in and obtain an access token.
IP access restrictions is a capability that controls the source network of connections, and it does not control whether individual users or groups can use the app.
RBAC is a mechanism for granting management permissions on Azure resources, and it does not apply to controlling app usage as required here.
Authentication and authorization in Azure App Service and Azure Functions
Restrict a Microsoft Entra app to a set of users
Q6.Your organization operates Microsoft Security Copilot.
Currently, contributors in Security Copilot can not only create custom plugins for their own sessions but also manage custom plugins for the entire organization.
You must prevent contributors from managing organization-wide custom plugins.
However, this must not affect the ability of contributors to create custom plugins for their own sessions.
What should you select in the plugin settings?
Show answer
When you set the users who can manage tenant-scope custom plugins to owners only, contributors can no longer add or manage plugins used across the organization.
Meanwhile, the user-scope setting is not changed, so contributors can continue to create and manage custom plugins for use in their own sessions.
Option B restricts even user-scope permissions, which violates the requirement.
Option A continues to allow contributors to manage plugins organization-wide, and option C specifies only user scope, so it cannot restrict tenant-wide management permissions.
In official Microsoft documentation, the product name Microsoft Security Copilot is still in use.
Manage plugins in Microsoft Security Copilot
Manage plugins and agents in Microsoft Security Copilot
Q7.You have an Azure Container Instances container group named CG1 with a DNS name of cg1.contoso.com.
CG1 is configured as follows.
A Linux container named container1 that serves HTTPS on TCP port 443 and hosts an application named App1.
A Linux container named container2 that listens on TCP port 5000 and is accessed only by App1.
A public IP address.
A security review has found that external clients can reach TCP port 5000 through the public IP address of CG1.
You must meet the following requirements.
Allow external clients to access container1 using only TCP port 443.
Allow container1 to continue accessing container2.
What should you configure?
Select the appropriate options in the answer area.
Show answer
To prevent external clients from connecting to port 5000, expose only port 443 on the public IP address of the container group.
Containers within the same Azure Container Instances container group share the same host and network namespace, so they can communicate with each other over localhost even on ports that are not exposed externally.
Therefore, specify localhost:5000 as the endpoint for App1 to connect to container2.
Using cg1.contoso.com:5000 would require exposing the port on the public side, which fails to meet the requirement of preventing external access.
Azure Container Instances is still the official service name.
Container groups in Azure Container Instances
Tutorial: Deploy a multi-container group using a Resource Manager template
Q8.An organization plans to make Azure OpenAI resources available to developers.
The management team wants the permissions granted to follow the principle of least privilege.
What should you use?
Show answer
With Azure RBAC, you can assign to users, groups, service principals, and others only the permissions needed, by specifying resources and scopes.
Therefore, granting the minimum built-in or custom role that corresponds to the developer’s work achieves least privilege.
Azure Policy exclusions define what is excluded from policy enforcement, and are not a mechanism for granting access rights.
Resource locks prevent accidental deletion or changes, and management groups hierarchically organize subscriptions, but neither controls individual access permissions.
Note that the latest Microsoft Learn uses Microsoft Foundry as the name of a related platform, and Azure RBAC continues to be used for individual access management to Azure OpenAI.
Role-based access control for Azure OpenAI Service
Azure built-in roles
Q9.A company wants to grant Just-In-Time access to privileged Azure roles instead of assigning permissions to administrators permanently.
Which Microsoft Entra feature should you implement?
Show answer
With Microsoft Entra Privileged Identity Management (PIM), you can register administrators as eligible users for Azure roles and have them activate a role for a limited period only when needed.
The main benefit is that it reduces standing privileges and grants access rights Just-In-Time for only the time needed.
Also, at activation you can require multifactor authentication, approval, and a justification to strictly control the use of privileges.
Dynamic groups are for attribute-based membership management, access packages are for requesting access to resources and lifecycle management, and self-service password reset is for account recovery, so they are not suited to this requirement.
The product name is still Microsoft Entra Privileged Identity Management.
What is Microsoft Entra Privileged Identity Management?
Eligible and time-bound role assignments in Azure RBAC
Q10.You have an Azure Storage account named storage1 that contains an Azure file share.
An application named App1 accesses this file share by using a system-assigned managed identity.
Meanwhile, administrators access the file share by using the storage account access keys.
You must allow App1 to access the file share without using the storage account access keys.
What should you configure on storage1?
Show answer
For App1 to read Azure Files without holding an access key, use the managed identity as a security principal in Microsoft Entra ID and grant an Azure RBAC data access role.
Assigning the Storage File Data Privileged Reader role to the managed identity of App1 enables read access using OAuth.
Option A continues to use keys, and option B only disables key authentication without granting permissions.
Option C is a setting that changes the default authentication method in the Azure portal, and it does not grant App1 access rights.
Note that the current name of Azure AD is Microsoft Entra ID.
Assign share-level permissions to an Azure file share