SC-500 · Domain 4: Manage and monitor security posture · 23% of the exam
Task 4.1: Manage security posture by using Defender for Cloud
Defender CSPM risks, compliance against frameworks, workload protection plans, AWS and GCP connectors, Defender Vulnerability Management settings for Azure VMs, and Defender EASM.
Study it
Defender CSPM, attack paths and security frameworks
Lesson coming
Workload protection plans and AWS and GCP connectors
Lesson coming
Defender Vulnerability Management and Defender EASM
Lesson coming
Sample questions
Try each one before opening the answer. Every option is explained, with the Microsoft documentation page that proves it.
Question 1 · choose 1
Your AWS organization has 120 member accounts, and the cloud team wants Defender for Cloud to cover all of them, including accounts created later. The team proposes connecting through its security tooling account, which is a delegated administrator for AWS Organizations, because nobody wants to use the management account. What should you tell them?
AConnect each of the 120 member accounts individually, which is the only way to cover an organization
BConnect the delegated administrator account instead of the management account, which then provisions all member accounts
CConnect any one member account and enable organization discovery in its connector
DConnect the AWS management account, which auto-provisions connectors for current and future member accounts
Show the answer and why
AConnect each of the 120 member accounts individually, which is the only way to cover an organization
Incorrect
Management account onboarding covers member accounts and new ones automatically.
BConnect the delegated administrator account instead of the management account, which then provisions all member accounts
Incorrect
Delegated administrator accounts aren't supported for organization onboarding.
CConnect any one member account and enable organization discovery in its connector
Incorrect
A member account connector covers only that account.
DConnect the AWS management account, which auto-provisions connectors for current and future member accounts
Correct
Only the management account can onboard the organization.
To onboard an AWS organization, create a connector for the AWS management account; delegated administrator accounts aren't supported. Auto provisioning then creates connectors for discovered member accounts and for accounts created later. The management account deploys a CloudFormation StackSet, which needs trusted access for AWS Organizations.
After discovery runs in Defender EASM, hundreds of hosts land in the Candidate state. Their findings don't appear on the dashboards, and they aren't rescanned daily. The team confirmed that 40 of them belong to the company. What should it do with those 40?
ALabel them Monitor Only
BLabel them Dependency
CAdd them as new discovery seeds and leave their state unchanged
DChange their state to Approved Inventory
Show the answer and why
ALabel them Monitor Only
Incorrect
Monitor Only is for related assets that the company doesn't control.
BLabel them Dependency
Incorrect
Dependency is for third-party infrastructure that supports your assets.
CAdd them as new discovery seeds and leave their state unchanged
Incorrect
Seeds guide discovery but don't change how existing assets are treated.
DChange their state to Approved Inventory
Correct
Approved Inventory assets are scanned daily and appear in dashboard charts by default.
Defender EASM asset states drive how assets are processed. Approved Inventory is your owned attack surface: always included in dashboard charts and scanned daily. Candidate and Requires Investigation assets need manual review of ownership, while Dependency and Monitor Only separate third-party and related assets for reporting.
A 30-day Defender Vulnerability Management exception for a recommendation on a group of servers is about to expire, and the vendor fix is delayed two more weeks. What should you do?
AExtend the existing exception by 14 days
BCreate a new exception with the needed duration and scope when the current one ends
CCancel the exception, which converts it to a permanent one
DMove the servers to a device group without recommendations
Show the answer and why
AExtend the existing exception by 14 days
Incorrect
Exception durations can't be extended.
BCreate a new exception with the needed duration and scope when the current one ends
Correct
A new exception is the documented way to continue past expiry.
CCancel the exception, which converts it to a permanent one
Incorrect
Cancelling ends the exception.
DMove the servers to a device group without recommendations
Incorrect
Recommendations follow the software, not an empty device group.
In Defender Vulnerability Management, an exception for a recommendation or CVE has a justification, a duration and a scope (global or by device group). Exception durations can't be extended; if an exception must continue after it expires, create a new one with the required duration and scope.
A zero-day affects one version of a PDF tool on servers covered by Defender for Servers Plan 2. Patching will take days. Security wants that version prevented from running meanwhile, while the patched version keeps working. What should you do?
AUse Block from the tool's security recommendation in Defender Vulnerability Management
BUse Warn from the recommendation
CCreate a recommendation exception until the patch is ready
DUninstall the tool with a Defender for Endpoint live response script on each server
Show the answer and why
AUse Block from the tool's security recommendation in Defender Vulnerability Management
Correct
Blocking creates file indicators for the vulnerable version's executables, which Defender Antivirus blocks.
BUse Warn from the recommendation
Incorrect
The requirement is to stop the version from running, not to discourage it.
CCreate a recommendation exception until the patch is ready
Incorrect
An exception hides the finding; it doesn't stop the app from running.
DUninstall the tool with a Defender for Endpoint live response script on each server
Incorrect
Removing the tool also removes the patched version's functionality and isn't the built-in mitigation.
From a security recommendation, Defender Vulnerability Management can block or warn on vulnerable versions of an app as a temporary mitigation. Block creates file-hash indicators for each executable in those versions, and Microsoft Defender Antivirus blocks them, usually within minutes but up to three hours. It's a premium capability, included in Defender for Servers Plan 2.
Defender for App Service was enabled on a subscription this week. Last month, before that, a team deleted a Linux web app but left its custom domain's CNAME record at an external registrar, pointing to the app's azurewebsites.net name. No dangling DNS alert has appeared. What should you do?
AMove the domain to Azure DNS, which dangling DNS detection requires
BRecreate the app on a Windows plan, which detection covers
CWait for the plan's first scan of the registrar's existing records
DDelete the leftover record yourself; no alert will come for it
Show the answer and why
AMove the domain to Azure DNS, which dangling DNS detection requires
Incorrect
Detection works with Azure DNS and with external registrars.
BRecreate the app on a Windows plan, which detection covers
Incorrect
Dangling DNS protection applies to App Service on both Windows and Linux.
CWait for the plan's first scan of the registrar's existing records
Incorrect
Defender for Cloud doesn't scan your registrar for existing dangling entries.
DDelete the leftover record yourself; no alert will come for it
Correct
The plan alerts when a site it has monitored is decommissioned with its DNS entry left; a site deleted before the plan was on won't be flagged.
Defender for App Service identifies DNS entries left at a registrar when an App Service website is decommissioned, which are vulnerable to subdomain takeover. It doesn't scan the registrar for existing dangling entries: it alerts when a site is decommissioned and its custom domain isn't deleted. It might not alert if the custom domain doesn't point directly to an App Service resource or if it hasn't monitored traffic to the site since the protection was enabled. It works with Azure DNS or external registrars, on Windows and Linux.
Each time the recommendation Storage account public access should be disallowed becomes unhealthy for a resource, a Logic App must open a ticket for the resource owner. What should you configure in Defender for Cloud?
AA workflow automation triggered by that recommendation
BContinuous export of the recommendation to an event hub
CAn Azure Policy remediation task
DA governance rule with an email notification only
Show the answer and why
AA workflow automation triggered by that recommendation
Correct
Workflow automations run Logic Apps when recommendations, alerts or compliance results change.
BContinuous export of the recommendation to an event hub
Incorrect
Exporting sends data but doesn't start the ticket workflow by itself.
CAn Azure Policy remediation task
Incorrect
Remediation tasks fix resources; they don't open tickets.
DA governance rule with an email notification only
Incorrect
Email notifications don't open a ticket in the ticketing system.
Workflow automation in Defender for Cloud triggers Azure Logic Apps when security alerts, recommendations or regulatory compliance assessments change, so that you can automate responses such as opening tickets or notifying owners.
Which scan intervals can you choose for a Defender for Cloud AWS or GCP connector?
AEvery 15 minutes or every hour
BOnly every 24 hours
CEvery 4, 6, 12 or 24 hours
DOnce a week
Show the answer and why
AEvery 15 minutes or every hour
Incorrect
The options are longer.
BOnly every 24 hours
Incorrect
You can pick shorter intervals too.
CEvery 4, 6, 12 or 24 hours
Correct
These are the selectable intervals; some collectors run on fixed schedules.
DOnce a week
Incorrect
Weekly isn't an option.
When you create an AWS or GCP connector, you select an interval to scan the environment every 4, 6, 12 or 24 hours. Some data collectors run with fixed intervals that the setting doesn't affect.
MDVM now sits under Exposure Management for cloud and device environments.
BUnder Endpoints > Vulnerability management, as before
Incorrect
That was the earlier location; it moved under Exposure Management.
CUnder Settings > Microsoft Sentinel
Incorrect
That configures Sentinel, not vulnerability management.
DOnly in the Azure portal
Incorrect
It's in the Defender portal.
Microsoft Defender Vulnerability Management is now under Exposure Management and supports both cloud and device environments; vulnerability scanning results are at Exposure Management > Vulnerability Management in the Defender portal.
Which risk levels does Defender for Cloud assign to recommendations with Defender CSPM?
ASevere, Moderate and Minor
BCritical, High, Medium, Low and Not evaluated
CRed, Amber and Green
D1 to 10
Show the answer and why
ASevere, Moderate and Minor
Incorrect
Those aren't the risk level names.
BCritical, High, Medium, Low and Not evaluated
Correct
These are the documented risk level classifications.
CRed, Amber and Green
Incorrect
Risk levels aren't colors.
D1 to 10
Incorrect
Risk levels aren't numeric scores.
Defender for Cloud's risk prioritization classifies recommendations as Critical, High, Medium, Low or Not evaluated, based on risk factors such as exploitability and business impact.
What Azure permission do you need to create a Defender for Cloud connector for a GCP project?
AReader on the subscription
BContributor on the Azure subscription
CGlobal Administrator in Microsoft Entra ID
DSecurity Reader on the subscription
Show the answer and why
AReader on the subscription
Incorrect
Reader can't create the connector.
BContributor on the Azure subscription
Correct
Contributor-level permission on the subscription is required.
CGlobal Administrator in Microsoft Entra ID
Incorrect
An Entra role isn't the requirement.
DSecurity Reader on the subscription
Incorrect
Security Reader is read-only.
To connect a GCP project or organization, you need Defender for Cloud enabled on the Azure subscription, Contributor-level permission on that subscription, and access to the GCP project or organization.