SC-500 replaces AZ-500 (retiring August 31, 2026) and carries forward nearly all of its identity, networking, storage/database, and Defender for Cloud/Sentinel scope — but adds an entire AI-security sub-domain covering Microsoft Copilot, Microsoft Foundry agents, and Microsoft Entra Agent ID that AZ-500 never tested. If you're coming from an AZ-500 background, expect most of this to feel familiar except questions 12 and 13 below.
Free practice material for SC-500 is thin right now because the exam is brand new — most generic “Azure security” question sets in circulation are just recycled AZ-500 content that never mentions Entra Agent ID, Purview DSPM, or Foundry AI Gateway. The 20 questions below are built directly against SC-500's four official domains, weighted the same way the real exam is, with the largest share on storage/database/network security since that domain is 25–30%.
Read the scenario, pick an answer, then click Reveal answer to check your reasoning against the explanation — not just the correct letter. Treat a wrong answer as a signal to review that specific domain in the full SC-500 practice exam rather than moving on.
Question distribution by domain
Manage identity, access, and governance
Q1–5 · 20–25% of examA finance analyst occasionally needs Owner access on a resource group to run a monthly reconciliation task, but currently holds a permanent, standing Owner assignment used only a few hours a month.
What should you configure to reduce this standing risk?
- A.Make the assignment eligible in Privileged Identity Management, requiring activation with justification and a time limit
- B.Downgrade the assignment to Contributor, which is broad enough for most tasks
- C.Leave the assignment as-is but enable Azure AD sign-in logging
- D.Create a scheduled Logic App that removes and re-adds the role monthly
Reveal answer & explanationHide answer
Correct: A. PIM eligible assignments remove standing access entirely — the role only becomes active after a time-boxed, justified activation, directly reducing the window a privileged role is usable. Downgrading to Contributor still leaves a standing broad assignment. Logging observes risk without reducing it. A custom Logic App reimplements what PIM already provides natively and less reliably.
Security wants to automatically require MFA — not block outright — whenever a sign-in is flagged as medium risk by Entra ID Protection, while low-risk sign-ins proceed normally.
What should you configure?
- A.A conditional access policy with a sign-in risk condition set to medium-and-above, requiring MFA as the grant control
- B.A conditional access policy that blocks all sign-ins outside the corporate network
- C.An Azure Policy definition auditing sign-in risk scores
- D.A Key Vault access policy restricting secret retrieval to compliant devices
Reveal answer & explanationHide answer
Correct: A. Conditional access supports sign-in risk as a condition, with MFA (rather than block) as the grant control — exactly the graduated response described. A network-location block is unrelated to risk scoring. Azure Policy governs resource configuration, not sign-in behavior. A Key Vault access policy is scoped to vault access, not general sign-in risk.
An Azure Function needs to retrieve a secret from Key Vault at runtime. The security team wants no credential to store, rotate, or risk leaking in application settings.
What should you implement?
- A.A system-assigned managed identity on the Function, granted Key Vault Get permission
- B.A service principal with a client secret stored in Function app settings
- C.A shared access signature token embedded in the Function code
- D.The Key Vault's own master key, retrieved once and cached indefinitely
Reveal answer & explanationHide answer
Correct: A. A managed identity is Azure-managed with no credential to store or rotate at all — the Function authenticates natively, and Key Vault RBAC/access policy grants it Get on the specific secret. A service principal secret still requires storage and rotation. A SAS token embedded in code is a leak risk. Key Vault does not expose a single "master key" for application retrieval.
A Key Vault must be reachable only from a specific virtual network and from trusted Azure services (like Azure Backup), with all other public internet access denied.
What should you configure?
- A.Key Vault firewall rules scoped to the VNet, with "Allow trusted Microsoft services" enabled
- B.An RBAC role assignment restricting Get permission to the backup service
- C.A conditional access policy blocking public IP ranges
- D.A Defender for Key Vault alert rule for public access attempts
Reveal answer & explanationHide answer
Correct: A. Key Vault firewall settings are the network-layer control for restricting reachability to specific VNets/IPs while still allowing the "trusted Microsoft services" exception for first-party services like Backup. RBAC governs who/what can act on secrets once connected, not network reachability. Conditional access does not apply to resource-to-resource networking. Defender for Key Vault detects after the fact — it does not restrict access.
An access review reveals a contractor's account holds Owner on a production subscription but has only ever used it to read resource configurations and never modified anything.
What is the correct remediation?
- A.Reassign the account to a built-in Reader role, or a custom role scoped to only the read actions actually used
- B.Leave the assignment unchanged since no misuse has occurred yet
- C.Disable the account entirely, even though the contractor still needs read access
- D.Move the account to a separate Entra tenant to isolate the risk
Reveal answer & explanationHide answer
Correct: A. RBAC evaluation and remediation of overprivileged access means matching the assignment to actual usage — Reader (or a scoped custom role) satisfies the observed need while removing standing write/delete capability. "No misuse yet" is not a defense against blast radius. Disabling the account breaks legitimate access the contractor still needs. Moving tenants does not right-size a permission — it just relocates the same overprivileged assignment.
Secure storage, databases, and networking
Q6–11 · 25–30% of exam — the largest domainA storage account already restricts network access to a specific VNet via firewall rules, but the security team also wants to be alerted if a malicious file is uploaded to a container inside it.
What should you additionally enable?
- A.Defender for Storage threat protection on the storage account
- B.A stricter storage firewall rule with a smaller IP range
- C.A resource lock preventing deletion of the storage account
- D.An Azure Policy audit rule checking firewall configuration
Reveal answer & explanationHide answer
Correct: A. The firewall rule is a prevention control (who can reach the account); malware/anomaly detection on uploaded content is a separate, detection-layer capability provided by Defender for Storage. Tightening the firewall further does not add malware scanning. A resource lock prevents deletion, unrelated to content threats. A policy audit checks configuration, not file content.
A compliance requirement mandates that every query executed against an Azure SQL Database be logged for later review, and that anomalous query patterns (e.g., a potential SQL injection attempt) trigger a security alert.
What combination satisfies both requirements?
- A.Enable Azure SQL Database auditing for the query log, and Defender for Databases for anomaly-based threat alerts
- B.Enable Defender for Databases alone, since it also stores a full query log
- C.Enable auditing alone, and manually review logs weekly for anomalies
- D.Configure a firewall rule restricting the database to the application's subnet only
Reveal answer & explanationHide answer
Correct: A. Auditing and Defender for Databases are complementary, not redundant: auditing produces the durable query log compliance requires, while Defender for Databases adds anomaly-based threat detection (like SQL injection patterns) auditing alone does not generate. Manual weekly review misses time-sensitive threats. A firewall rule restricts network reachability, not query-level auditing or threat detection.
A three-tier application has web, application, and database VMs whose IP addresses change as the environment autoscales. You need firewall rules that reference the VM's role rather than its current IP address.
What should you implement?
- A.Application security groups (ASGs) grouping VMs by role, referenced inside NSG rules instead of individual IPs
- B.A separate NSG per individual VM, updated manually as IPs change
- C.A single NSG rule allowing all traffic between the three subnets
- D.Azure Firewall rules referencing each VM's current IP, refreshed nightly
Reveal answer & explanationHide answer
Correct: A. ASGs let NSG rules reference a logical group (e.g., "database-tier") instead of specific IPs, so autoscaling or IP changes don't require rule updates. Per-VM NSGs and nightly IP refreshes both reintroduce the exact manual-maintenance problem ASGs solve. Allowing all subnet-to-subnet traffic abandons segmentation entirely.
An application inside a virtual network must connect to an Azure SQL Database entirely over the Microsoft backbone, with no traffic ever traversing the public internet, and the database should not be reachable from outside that VNet at all.
What should you configure?
- A.A private endpoint for the Azure SQL Database, connected into the application's virtual network
- B.A service endpoint on the SQL server's public firewall rules
- C.A Private Link service exposing the application to other tenants
- D.Microsoft Entra Private Access configured for the application's users
Reveal answer & explanationHide answer
Correct: A. A private endpoint brings the PaaS resource's (SQL Database's) private IP directly into the consuming VNet, keeping traffic off the public internet and restricting reachability to that private connection. A service endpoint still routes over Microsoft's backbone but keeps the resource's public endpoint reachable via firewall rule, not private-IP-only. A Private Link service is for exposing your own service outward, the opposite direction. Entra Private Access secures user access to apps, not application-to-database connectivity.
Remote employees need access to an internal, on-premises line-of-business application without a traditional site-to-site VPN client, and access should be governed by conditional access policies per application.
What should you implement?
- A.Microsoft Entra Private Access, publishing the application and enforcing conditional access per app
- B.A private endpoint connecting the on-premises app into an Azure VNet
- C.Azure Virtual Network Manager network access policies
- D.A Private Link service exposing the on-prem app to other Azure tenants
Reveal answer & explanationHide answer
Correct: A. Entra Private Access is purpose-built for secure, VPN-less user access to private/on-prem applications, with per-application conditional access — exactly this scenario. A private endpoint secures PaaS-to-VNet connectivity, not user-to-on-prem-app access. Virtual Network Manager governs network topology/access policy between Azure networks, not remote user access. A Private Link service exposes a service to other tenants, not to remote employees.
A VM's NSG has an explicit "Allow" rule for port 443 from a specific subnet, but the application team reports the traffic is still being blocked. Both the subnet NSG and the NIC NSG could be involved.
What should you use to determine the actual effective rule being applied?
- A.Network Watcher's effective security rules diagnostic on the VM's network interface
- B.A packet capture on the destination VM only
- C.The Azure Activity Log for the NSG resource
- D.Azure Policy compliance state for the NSG
Reveal answer & explanationHide answer
Correct: A. Effective security rules merges every NSG (subnet-level and NIC-level) applying to a given NIC and shows the actual resulting rule set — the direct way to find a conflicting deny rule elsewhere in the chain. A packet capture shows traffic outcomes but not which rule caused a block. The Activity Log records management operations, not rule evaluation results. Azure Policy compliance state checks configuration against a policy definition, not live traffic evaluation.
Secure compute
Q12–16 · 20–25% of examAn organization deployed a Microsoft Foundry agent with a Microsoft Entra Agent ID identity that accesses sensitive financial data. Only the finance security group should be able to invoke it — all other employees must be denied, without disabling any other application access.
What should you configure?
- A.A conditional access policy scoped to the Entra Agent ID, granting access only to the finance security group
- B.A Microsoft Purview DLP policy blocking financial data from leaving the tenant
- C.An Azure Policy assignment restricting the agent's resource group
- D.A Defender for Cloud workload protection plan for the agent's hosting resource
Reveal answer & explanationHide answer
Correct: A. Conditional access can target an Entra Agent ID directly, restricting which users/groups can invoke that specific agent while leaving every other application untouched. Purview DLP governs data exfiltration, not who can invoke the agent. Azure Policy enforces resource configuration, not per-agent user access. A Defender for Cloud workload protection plan detects threats after deployment — it does not gate invocation.
Before broadly enabling Microsoft Copilot, security wants to identify SharePoint sites and documents that are more broadly shared than necessary and could be surfaced by Copilot to users who should not see them.
What should you use?
- A.Microsoft Purview Data Security Posture Management (DSPM) for AI
- B.Defender for Cloud regulatory compliance dashboard
- C.Microsoft Entra ID access reviews on SharePoint site owners
- D.Azure Policy audit rules for SharePoint Online
Reveal answer & explanationHide answer
Correct: A. Purview DSPM for AI is purpose-built to identify overexposed data (like broadly-shared SharePoint content) before Copilot or other AI apps can surface it inappropriately. The Defender for Cloud compliance dashboard covers Azure resource compliance, not SharePoint sharing posture. Entra access reviews recertify group/role membership, not document-level sharing exposure. Azure Policy does not govern SharePoint Online content.
Administrators currently RDP into VMs using a permanently open port 3389, with a public IP assigned to each VM. Security wants to eliminate the public IP exposure entirely, not just shorten how long the port stays open.
What should you implement?
- A.Azure Bastion, removing the need for a public IP or open RDP/SSH port on the VMs at all
- B.Just-in-time VM access, which still requires a public IP but limits the exposure window
- C.A stricter NSG rule allowing RDP only from the office IP range
- D.Azure Firewall DNAT rules forwarding RDP from a single public IP
Reveal answer & explanationHide answer
Correct: A. Bastion provides browser-based RDP/SSH through the Azure portal without any public IP or open management port on the VM itself — it solves "eliminate public exposure," not just "shorten the window." JIT still assumes an accessible port that opens temporarily. A tighter NSG rule and a Firewall DNAT rule both still expose a port, just to a narrower source.
A team deploys new workloads to AKS frequently and wants automatic detection of insecure configurations (like privileged containers or missing resource limits) and known vulnerabilities in container images, without manually reviewing every manifest.
What should you enable?
- A.Defender for Containers on the AKS cluster
- B.Azure Policy audit mode only, with no enforcement
- C.Network Watcher flow logs on the AKS subnet
- D.Defender for Storage on the container registry's backing storage account
Reveal answer & explanationHide answer
Correct: A. Defender for Containers is purpose-built to detect runtime misconfigurations and image vulnerabilities across AKS, ACR, and container workloads generally. Azure Policy audit mode reports policy violations but is not a vulnerability/misconfiguration scanner in the same sense. Network Watcher flow logs capture network traffic, not container configuration risk. Defender for Storage protects storage accounts, not container image content.
An Azure Function exposes a back-end API that should only be callable through a gateway enforcing subscription-key authentication and per-client rate limiting, without modifying the Function's own code.
What should you configure?
- A.Azure API Management in front of the Function, with subscription-key and rate-limit policies applied at the gateway
- B.A conditional access policy on the Function App's Entra ID application registration
- C.A Web Application Firewall policy on Azure Front Door
- D.Azure Firewall network rules restricting inbound traffic to the Function's IP
Reveal answer & explanationHide answer
Correct: A. API Management sits in front of a back-end API and enforces policies like subscription keys and rate limiting declaratively at the gateway, with no back-end code changes. Conditional access governs user/app sign-in, not per-call API rate limiting. A WAF protects against web exploits (e.g., injection) but does not natively enforce subscription-key auth or granular rate limits. Azure Firewall network rules operate at IP/port level, not API-key or rate-limit granularity.
Manage and monitor security posture
Q17–20 · 20–25% of examA security team believes some subdomains and cloud resources were created outside standard provisioning processes over the years and are not tracked in the organization's current asset inventory.
What should you use to discover them?
- A.Microsoft Defender External Attack Surface Management (EASM)
- B.Defender Cloud Security Posture Management (CSPM)
- C.Microsoft Defender Vulnerability Management
- D.Azure Resource Graph queries against the current subscription
Reveal answer & explanationHide answer
Correct: A. Defender EASM is specifically built to discover unmanaged, shadow-IT, or externally exposed assets that are not already in your known inventory. CSPM evaluates posture on resources you already know about. Vulnerability Management scans known Azure VMs for software vulnerabilities. Resource Graph only queries resources already registered in your subscriptions — it cannot find what was never tracked there.
Whenever Microsoft Sentinel creates a high-severity incident involving a specific alert type, the team wants an existing Logic App that disables the affected user account to run automatically, without an analyst manually triggering it.
What should you configure?
- A.An automation rule that triggers on the matching incident conditions and runs the existing playbook
- B.The playbook alone, scheduled to run every five minutes regardless of incidents
- C.A Sentinel workbook displaying the incident for manual action
- D.A custom log table capturing the incident for later batch review
Reveal answer & explanationHide answer
Correct: A. An automation rule is the orchestration layer that reacts to matching incident conditions and invokes a playbook (the Logic App) to take the remediation action — exactly the "automatically, without manual triggering" requirement. Scheduling the playbook independently of any incident means it can't react to a specific triggering condition. A workbook is a visualization, not an action. A custom log table stores data — it doesn't trigger a response.
An organization runs workloads in both Azure and AWS and wants unified security posture visibility and recommendations across both clouds from a single Defender for Cloud view.
What should you configure?
- A.Connect the AWS account to Defender for Cloud using its multicloud connector
- B.Deploy Defender for Servers only inside Azure and manually cross-reference AWS Security Hub separately
- C.Enable Defender CSPM in Azure only, since AWS resources cannot be evaluated by Defender for Cloud
- D.Migrate the AWS workloads into Azure to consolidate visibility
Reveal answer & explanationHide answer
Correct: A. Defender for Cloud has a native multicloud connector for AWS (and GCP) that pulls posture and recommendations into the same unified view as Azure resources. Manually cross-referencing two separate consoles defeats the purpose of unified visibility and doesn't scale. AWS resources absolutely can be evaluated once connected — CSPM is not Azure-only. Migrating workloads is a disproportionate, unnecessary response to a visibility requirement.
A SOC wants analysts to use Microsoft Security Copilot to investigate incidents, pulling in context from an additional third-party threat intelligence source beyond what Copilot integrates by default.
What should you configure?
- A.A Security Copilot workspace with the relevant plugin enabled for that data source
- B.A Sentinel data connector only, with no Security Copilot configuration
- C.A Defender for Cloud workload protection plan referencing the third-party source
- D.An Azure Policy initiative requiring analysts to check the third-party source manually
Reveal answer & explanationHide answer
Correct: A. Security Copilot extends its investigative capability through plugins configured within a workspace, letting it pull in additional data sources like third-party threat intelligence. A Sentinel data connector ingests data into Sentinel but does not by itself extend Security Copilot's plugin capabilities. A workload protection plan is unrelated to Copilot tooling. A manual policy requirement doesn't integrate the data — it just documents a process.
Want 480 More SC-500 Questions?
Full coverage of all four domains, including the AI-security sub-domain, calibrated harder than the real exam. Start with 40 free — no card required.
Start Free SC-500 Practice →