AZ-802 — Microsoft Certified: Windows Server Hybrid Administrator Associate Cheat Sheet
Compact AZ-802 Cheat sheet for Microsoft Windows Server hybrid administration: security, high availability, disaster recovery, migration, monitoring, and troubleshooting.
Use the tables for a quick pre-exam check. Expand a topic’s notes for explanations, examples, and additional distinctions.
Scope and study context
AZ-802 is not just “Windows Server in Azure.” It tests whether you can secure, migrate, protect, monitor, and troubleshoot Windows Server workloads across on-premises and hybrid environments. Expect scenario questions where the best answer depends on constraints such as downtime tolerance, identity model, data size, recovery objective, network connectivity, administrative scope, and whether the workload is physical, virtual, clustered, or cloud-connected.
AZ-802 Exam Focus at a Glance
This Cheat Sheet supports independent preparation for Microsoft AZ-802, Microsoft Certified: Windows Server Hybrid Administrator Associate (AZ-802). Use it to review high-yield decisions for Windows Server in hybrid environments: on-premises, Azure VMs, Azure Arc-enabled servers, identity, security, high availability, disaster recovery, migration, monitoring, and troubleshooting.
| If the scenario asks for… | Think first about… | Common trap |
|---|---|---|
| Manage non-Azure Windows Servers from Azure | Azure Arc-enabled servers | Arc is management, not automatic migration |
| Patch servers at scale | Azure Update Manager, maintenance configurations | Do not assume WSUS and Azure Update Manager are identical |
| Collect logs and performance data | Azure Monitor Agent, Data Collection Rules, Log Analytics | Agent installed but no DCR means little or no data |
| Improve security posture | Microsoft Defender for Cloud, Microsoft Defender for Endpoint, security baselines | Defender for Cloud posture management is not the same as antivirus |
| Back up files, system state, or VMs | Azure Backup / Recovery Services vault | Backup is not orchestrated disaster recovery |
| Replicate and fail over workloads | Azure Site Recovery | ASR is not long-term backup retention |
| Migrate servers to Azure | Azure Migrate | Assessment and replication/cutover are separate phases |
| Migrate file servers | Storage Migration Service, Azure File Sync, AzCopy/Robocopy depending target | File sync is not the same as one-time migration |
| Provide app or VM high availability | Failover clustering, NLB, load balancers, Storage Spaces Direct | HA inside a site is not a full DR strategy |
| Troubleshoot hybrid visibility | Arc agent, Azure Monitor Agent, DCRs, RBAC, network/proxy | “Server is online” does not mean Azure can manage it |
Hybrid Management Service Selection
| Tool or service | Best use | Key AZ-802 cues | Watch for |
|---|---|---|---|
| Windows Admin Center | Browser-based Windows Server, cluster, Hyper-V, and storage management | Administer servers without full RDP; integrate with Azure services | WAC is a management gateway, not a cloud control plane by itself |
| Server Manager | Traditional role/feature and remote server management | Small on-prem admin scenarios | Less useful for Azure-scale governance |
| Azure Arc-enabled servers | Project non-Azure Windows/Linux servers into Azure Resource Manager | On-premises or other-cloud servers need Azure Policy, tags, Defender, Update Manager, monitoring | Requires Connected Machine agent and outbound connectivity |
| Azure Policy with guest configuration | Audit or enforce machine configuration through Azure | Compliance checks across Azure and Arc-enabled servers | Policy assignment scope and remediation identity matter |
| Azure Update Manager | Assess and deploy OS updates across Azure VMs and Arc-enabled servers | Scheduled patching, update compliance, maintenance windows | Requires supported agent/configuration; not a replacement for every WSUS use case |
| Azure Automation | Runbooks, automation jobs, hybrid workers | Repeatable operational tasks across hybrid infrastructure | Automation account permissions and hybrid worker placement matter |
| Microsoft Defender for Cloud | Security posture, recommendations, regulatory-style compliance views, workload protection integration | “Secure score,” recommendations, server protection, Arc machines | Not the same as Windows Defender Firewall or Defender Antivirus |
| Microsoft Defender for Endpoint | Endpoint detection and response | Investigate suspicious activity, endpoint alerts, attack timeline | Licensing/onboarding method is scenario-dependent |
| Azure Monitor | Metrics, logs, alerts, dashboards, VM insights | Centralized monitoring and KQL analysis | Data appears only if collection is configured |
| Log Analytics workspace | Stores queryable monitoring/log data | KQL, log retention, alerts from logs | Workspace region, permissions, and DCR association can block visibility |
| Azure Monitor Agent | Modern monitoring agent for guest logs/performance | Data Collection Rules, Azure/Arc servers | Replaces many legacy collection patterns, but legacy agents may still appear in existing environments |
| Recovery Services vault | Azure Backup and Azure Site Recovery container | Backup policies, protected items, ASR replication items | Vault choice affects management boundary and recovery configuration |
| Azure Migrate | Discovery, assessment, dependency analysis, server migration | Move VMware, Hyper-V, physical, or other servers to Azure | Assessment readiness is not the same as completed migration |
| Storage Migration Service | File server inventory, transfer, and cutover | Preserve shares/security while moving to newer Windows Server or Azure VM | Not designed as a continuous file sync service |
| Azure File Sync | Centralize file shares in Azure Files with Windows Server cache | Branch file servers, cloud tiering, multi-site file access | Sync topology and conflict behavior matter |
| Azure Site Recovery | VM/workload replication, test failover, planned/unplanned failover | DR to Azure or secondary site | Does not replace backups or app-level consistency planning |
Notes and examples
Quick service-selection table
| Need | Likely tool or feature | Watch for this trap |
|---|---|---|
| Manage Windows Servers across on-premises, edge, and multicloud from Azure | Azure Arc-enabled servers | Arc enables management; it does not automatically migrate servers |
| Centralized monitoring and queries | Azure Monitor with Log Analytics | Diagnostic data must be collected before you can query it |
| Security recommendations and posture management | Microsoft Defender for Cloud | Recommendations depend on resource visibility, configuration, and plans enabled |
| Backup files, folders, system state, or VMs | Azure Backup | Backup is not the same as live disaster recovery failover |
| Replicate workloads for failover to Azure or another site | Azure Site Recovery | ASR is for recovery orchestration, not long-term backup retention |
| Migrate servers to Azure | Azure Migrate | Assessment, dependency analysis, and replication are separate phases |
| Migrate file servers and preserve shares/permissions | Storage Migration Service | Name cutover and identity/permission preservation are key details |
| Sync branch files with cloud tiering | Azure File Sync | It is not a replacement for a backup strategy |
| Local high availability for roles/VMs | Failover clustering | Cluster availability does not protect against all-site failure |
| Replicate volumes between servers or clusters | Storage Replica | Replication can copy corruption or deletion; still back up |
| Patch clustered workloads with reduced disruption | Cluster-Aware Updating | Nodes must drain and resume correctly |
| Manage servers through browser-based tooling | Windows Admin Center | WAC is a management tool, not a monitoring platform by itself |
Identity, Directory, and Access Control
AD DS and Microsoft Entra ID distinctions
| Component | Primary purpose | Choose when… | Exam caution |
|---|---|---|---|
| Active Directory Domain Services | Kerberos/NTLM domain auth, domain join, Group Policy, LDAP, computer accounts | Windows Server workloads depend on domain services | Microsoft Entra ID does not directly replace all AD DS features |
| Microsoft Entra ID | Cloud identity, OAuth/OIDC/SAML apps, Azure RBAC integration | Users need cloud app access, Azure portal access, conditional access | Entra users are not automatically domain users for legacy apps |
| Microsoft Entra Connect / Cloud Sync | Synchronize identities from AD DS to Entra ID | Hybrid identity required | Know sync direction and sign-in method implications |
| Microsoft Entra Domain Services | Managed domain services in Azure | Azure workloads need LDAP/Kerberos/NTLM without managing DCs | Not the same as extending your existing DCs into Azure |
| Domain controller in Azure VM | Extend existing AD DS into Azure | Azure workloads need full AD DS control and replication | Treat as a DC: DNS, sites, subnets, backup, security |
Notes and examples
AD DS operations to recognize
| Task | High-yield reference | Useful checks |
|---|---|---|
| Add a domain controller | Install AD DS role, promote server, configure DNS and site placement | DNS health, replication, time sync |
| Replace old domain controllers | Add new DCs, transfer FSMO roles, validate replication, demote old DCs | Do not simply shut down the last role holder |
| Manage replication topology | Use AD Sites and Services, site links, subnets | Incorrect subnet mapping causes wrong DC selection |
| Troubleshoot logon issues | Check DNS SRV records, secure channel, time skew, DC locator | nltest, dcdiag, repadmin, w32tm |
| Protect privileged accounts | Tiered admin model, Protected Users, PAWs, JEA, LAPS | Avoid using domain admin for routine server tasks |
| Service account management | gMSA where supported | gMSA requires domain support and correct host authorization |
| Restore deleted AD objects | AD Recycle Bin when enabled | Not a substitute for full system state backup |
| Back up domain controllers | System State / supported backup methods | Avoid unsupported snapshots or rollback patterns |
Privileged access decision table
| Requirement | Prefer | Why |
|---|---|---|
| Local administrator password rotation | Windows LAPS | Unique, rotated local admin passwords reduce lateral movement |
| Run limited PowerShell admin tasks | Just Enough Administration | Provides role-limited endpoints instead of full admin shell |
| Manage services securely | Group Managed Service Account | Automatic password management and SPN support |
| Temporary Azure privileged role | Microsoft Entra Privileged Identity Management | Time-bound elevation for cloud roles |
| Restrict credential exposure during remote admin | Credential Guard / Remote Credential Guard where applicable | Reduces credential theft risk |
| Delegate server management through WAC | Windows Admin Center role-based controls and gateway access | Centralizes browser-based server administration |
Windows Server Security Reference
| Control | Use for | Implementation clues | Common trap |
|---|---|---|---|
| Microsoft Defender Antivirus | Malware protection on Windows Server | Real-time protection, definitions, exclusions | Exclusions must be justified; do not broadly exclude system paths |
| Microsoft Defender for Endpoint | EDR, investigation, advanced threat detection | Onboarding package, security portal alerts | Antivirus status alone does not confirm EDR onboarding |
| Microsoft Defender for Cloud | Posture management and workload protection recommendations | Secure score, recommendations, Arc/Azure servers | Recommendations may require agent, extension, or plan configuration |
| Windows Defender Firewall | Host-level inbound/outbound filtering | Profiles: domain/private/public; rule scope | NSGs do not replace host firewall rules |
| BitLocker | Volume encryption | TPM, recovery keys, policy enforcement | Encryption protects data at rest, not live compromised sessions |
| Secure Boot / TPM / Secured-core | Boot integrity and hardware-rooted protections | Modern server hardware or Azure VM generation support | Availability depends on platform capabilities |
| Credential Guard | Protect derived credentials | Virtualization-based security | Can affect older auth/delegation patterns |
| SMB signing/encryption | Protect SMB integrity/confidentiality | File server and client settings | SMB encryption is not a backup or access-control substitute |
| TLS certificate management | Secure service endpoints | AD CS, public CA, certificate lifecycle | Expired certs break hybrid services and agents |
| JEA | Least-privilege PowerShell operations | Role capabilities and session configuration | Users still need a defined endpoint and permissions |
| Security baselines | Standardized hardening | Microsoft security baselines, GPO, Intune, policy | Test before broad enforcement |
| Shielded VMs / Host Guardian Service | Protect Hyper-V VMs from fabric admins | Guarded fabric, attestation, key protection | More complex than normal VM encryption |
Notes and examples
Fast security checks
Get-MpComputerStatus | Select-Object AMServiceEnabled,AntivirusEnabled,RealTimeProtectionEnabled
Get-SmbServerConfiguration |
Select-Object EnableSMB1Protocol,EncryptData,RejectUnencryptedAccess
auditpol /get /category:*
Get-LocalUser | Where-Object Enabled -eq $true
Get-LocalGroupMember Administrators
Identity and administrative access
AZ-802 security questions often reward least privilege and controlled administration over broad local admin access.
| Concept | Know this | Candidate mistake |
|---|---|---|
| Least privilege | Grant only the rights needed for the task, preferably through roles or delegated administration | Giving Domain Admin for routine server management |
| Just Enough Administration | PowerShell constrained endpoints can expose only approved commands | Thinking JEA is the same as ordinary remote PowerShell |
| Privileged Access Workstations | Use hardened admin workstations for privileged operations | Administering domain controllers from general-purpose workstations |
| Local Administrator Password Solution | Manages unique local admin passwords | Reusing one local administrator password across many servers |
| Group Managed Service Accounts | Service accounts with automatic password management | Using normal user accounts for services and manually rotating passwords |
| Credential Guard | Helps protect credentials from theft on supported systems | Assuming it replaces all endpoint hardening |
| Windows Defender Firewall | Host-based traffic control | Disabling the firewall to “fix” connectivity instead of allowing required traffic |
| Secure remote access | Prefer secured management paths and audited administrative access | Exposing RDP broadly to the internet |
Active Directory Domain Services hardening
For hybrid Windows Server environments, AD DS is often the most important security dependency.
High-yield reminders:
- Domain controllers should be patched, monitored, backed up, and isolated from general workloads.
- Do not install unnecessary server roles or applications on domain controllers.
- Use separate administrative accounts for privileged administration.
- Protect privileged groups such as Domain Admins, Enterprise Admins, and Schema Admins.
- Audit authentication failures, privilege use, account changes, and directory changes.
- Confirm time synchronization; Kerberos depends on time.
- Use secure DNS configuration because AD DS depends heavily on DNS.
- Use read-only domain controllers where appropriate for locations with lower physical security.
Common trap: a question describes a branch office with poor physical security and asks how to provide local authentication. A read-only domain controller may be better than placing a writable domain controller there.
Server hardening decision rules
| If the scenario says… | Think… |
|---|---|
| “Reduce attack surface” | Remove roles/features, close ports, apply baselines, enforce firewall rules |
| “Protect credentials on servers” | Credential Guard, LSASS protection, admin tiering, avoid interactive logons |
| “Delegate a narrow admin task” | JEA, RBAC where available, constrained PowerShell |
| “Secure local admin passwords” | LAPS-style local password management |
| “Protect data at rest” | BitLocker, EFS where suitable, storage encryption |
| “Protect SMB traffic” | SMB signing/encryption depending on confidentiality/integrity need |
| “Detect threats and get recommendations” | Defender for Cloud / Defender integrations |
| “Assess compliance against security baselines” | Security policy, baselines, Defender for Cloud recommendations |
Certificates and PKI
Know the difference between certificate problems and identity problems.
| Symptom | Likely area to check |
|---|---|
| TLS warning or service refuses secure connection | Certificate name, trust chain, expiration, EKU, private key |
| Smart card or certificate logon fails | Certificate template, mapping, revocation, domain trust, time |
| Enrollment fails | Template permissions, autoenrollment policy, CA availability |
| Revocation check fails | CRL/OCSP publication and reachability |
| Internal service works on LAN but not externally | Subject/SAN, trust chain, firewall, DNS, certificate binding |
Common trap: renewing a certificate does not automatically update every application binding. The service may still be using the old certificate.
Networking and Remote Administration
| Requirement | Use | Notes |
|---|---|---|
| Secure server management without broad RDP exposure | Windows Admin Center, PowerShell Remoting, JEA | Prefer constrained, audited admin paths |
| Connect on-premises network to Azure | Site-to-site VPN or ExpressRoute | VPN is internet-based encrypted tunnel; ExpressRoute is private connectivity through provider |
| Connect one server to Azure VNet for management/testing | Azure Network Adapter through WAC, where suitable | Good for limited scenarios, not enterprise WAN design |
| Protect Azure VM traffic | NSG, Azure Firewall, route tables, host firewall | NSG filters at subnet/NIC; host firewall still matters |
| Protect on-prem server traffic | Windows Defender Firewall, network firewalls, IPsec | Azure controls do not automatically protect on-prem paths |
| Diagnose Azure network path | Network Watcher, Connection Monitor, effective routes/NSGs | Applies to Azure resources and monitored endpoints depending configuration |
| Remote command execution | WinRM / PowerShell Remoting | Requires listener, firewall, auth, and endpoint permissions |
| Remote GUI access | RDP, Azure Bastion for Azure VMs | Bastion is for Azure VM access, not general on-prem RDP |
Notes and examples
Common hybrid connectivity traps
| Symptom | Likely area to inspect |
|---|---|
| Azure Arc server disconnected | Outbound HTTPS/proxy, Connected Machine agent, identity/RBAC |
| Azure Monitor no data | DCR association, AMA health, workspace permissions, collection rule scope |
| Domain logons slow in Azure | AD Sites and Services subnet mapping, DNS, DC placement |
| Azure VM cannot join domain | DNS points to AD DS DNS servers, network path to DCs, time sync |
| WAC cannot manage server | WinRM, firewall rules, trusted hosts/domain trust, gateway permissions |
| Backup/ASR agent cannot register | Vault credentials, outbound connectivity, clock, proxy/TLS inspection |
Common AZ-802 traps
| Trap | Better thinking |
|---|---|
| Backup and replication are interchangeable | Backup restores previous points; replication supports failover or copy continuity |
| Azure Arc migrates servers | Arc manages and governs hybrid servers; migration is separate |
| Azure File Sync is backup | Sync can propagate deletions and corruption |
| Failover clustering protects against all disasters | It protects against certain local failures, not every site-wide dependency |
| Storage Replica removes the need for backup | Replication can replicate bad changes |
| Disabling firewalls is an acceptable fix | Create precise rules and verify required ports |
| Domain Admin is needed for routine tasks | Use delegation, JEA, and least privilege |
| A certificate renewal fixes all TLS issues | Bindings, trust chains, SANs, and private keys still matter |
| Monitoring starts after an incident | Telemetry must be collected before useful historical analysis |
| A successful migration means the app works | App dependencies, identity, DNS, and performance still require validation |
High Availability Reference
HA technology selection
| Scenario | Prefer | Why | Avoid assuming |
|---|---|---|---|
| Stateful workload needs automatic failover between nodes | Failover clustering | Cluster service coordinates resource ownership | Cluster alone provides site DR |
| Stateless scale-out TCP/UDP application | Network Load Balancing or external load balancer | Distributes client traffic | NLB protects shared state |
| Highly available Hyper-V storage | Cluster Shared Volumes, Storage Spaces Direct, SAN-backed cluster | Shared or replicated storage for clustered VMs | Local disks alone are enough |
| Highly available SMB application shares | Scale-Out File Server where appropriate | Active-active SMB access for application data | General user file shares always fit SOFS |
| Rolling patching of clusters | Cluster-Aware Updating | Coordinates node maintenance | Manual patching is always safe |
| Site-aware cluster | Failover cluster with site awareness, proper quorum/witness | Supports planned placement and failover logic | It eliminates need for DR testing |
| VM-level replica between hosts/sites | Hyper-V Replica | Asynchronous VM replication | Same as backup or app-aware HA |
| Volume-level replication | Storage Replica | Block-level replication between servers/clusters | Same as DFS Replication |
Notes and examples
Failover clustering quick checks
| Area | What to remember |
|---|---|
| Validation | Run cluster validation before creating or changing a supported cluster |
| Quorum | Prevents split-brain; witness helps maintain majority |
| Witness options | Disk witness, file share witness, cloud witness depending topology |
| Cloud witness | Useful when Azure is reachable and no shared witness disk is preferred |
| Dynamic quorum | Adjusts quorum vote behavior as nodes change |
| Cluster networks | Separate or logically plan client, storage, live migration, and management traffic where needed |
| CSV | Common for Hyper-V clustered VM storage |
| CAU | Automates patching workflow across cluster nodes |
| Drain roles | Move clustered roles before maintenance |
| Anti-affinity / preferred owners | Control workload placement patterns |
Cluster PowerShell snippets
Install-WindowsFeature Failover-Clustering -IncludeManagementTools
Test-Cluster -Node "SRV1","SRV2"
New-Cluster -Name "CL01" -Node "SRV1","SRV2" -StaticAddress "10.0.0.50"
Get-ClusterNode
Get-ClusterGroup
Get-ClusterQuorum
Cloud witness example pattern:
Set-ClusterQuorum -CloudWitness `
-AccountName "<storage-account-name>" `
-AccessKey "<storage-account-key>"
Failover clustering essentials
A failover cluster provides high availability for supported workloads by moving clustered roles between nodes.
Know these concepts:
| Concept | Meaning |
|---|---|
| Node | Server participating in the cluster |
| Clustered role | Workload managed by the cluster |
| Quorum | Voting mechanism that determines whether the cluster can continue running |
| Witness | Tie-breaker resource such as disk, file share, or cloud witness |
| CSV | Cluster Shared Volumes for shared access to storage by cluster nodes |
| Drain roles | Move workloads off a node before maintenance |
| Cluster-Aware Updating | Coordinates patching of cluster nodes while maintaining availability |
Quorum and witness logic
Avoid memorizing only one witness type. Understand the scenario.
| Witness type | Common fit |
|---|---|
| Cloud witness | Hybrid or multi-site environments with Azure connectivity |
| File share witness | Simple witness option when a reliable file share is available |
| Disk witness | Traditional shared-storage cluster scenarios |
| No witness | Certain configurations where node votes alone are appropriate |
Trap: a witness does not host the workload. It helps the cluster make quorum decisions.
High availability versus disaster recovery
| Requirement | Better fit |
|---|---|
| Survive a single host failure in the same datacenter | Failover clustering |
| Maintain app availability during node maintenance | Cluster-Aware Updating and role draining |
| Replicate a VM to another host/site for recovery | Hyper-V Replica or Azure Site Recovery, depending on scenario |
| Recover an entire site in Azure | Azure Site Recovery |
| Restore accidentally deleted or corrupted data | Backup |
| Keep two storage copies synchronized | Storage Replica |
Common mistake: choosing failover clustering for site disaster recovery without considering shared dependencies. A cluster may still fail if the entire site, network, storage, or identity dependency is unavailable.
Storage Spaces Direct and Storage Replica
| Feature | Primary purpose | Key exam angle |
|---|---|---|
| Storage Spaces Direct | Software-defined storage using local drives in clustered servers | High availability and scalable storage inside a cluster |
| Storage Replica | Block-level volume replication between servers or clusters | Disaster recovery or stretch-cluster storage replication |
| DFS Replication | File-level replication for certain file data scenarios | Not ideal for open files, databases, or low-RPO block replication |
Trap: Storage Replica is not a backup. If malware encrypts replicated data, the encrypted data may replicate too.
Disaster Recovery and Backup
Backup vs replication vs disaster recovery
| Requirement | Best fit | Reason |
|---|---|---|
| Restore accidentally deleted files | Azure Backup, Windows Server Backup, file backup | Point-in-time recovery |
| Restore Windows Server system state | Azure Backup with MARS agent or supported backup product | Protects critical OS roles such as AD DS |
| Long-term retention | Backup policy | Replication usually keeps only current or near-current state |
| Fail over VM workloads to Azure | Azure Site Recovery | Replication plus orchestration |
| Test failover without disrupting production | Azure Site Recovery test failover | Validates DR plan |
| Replicate storage volumes between servers/clusters | Storage Replica | Block-level volume replication |
| Replicate Hyper-V VMs between hosts | Hyper-V Replica | VM-focused asynchronous replication |
| Protect Azure VM | Azure Backup VM backup and/or ASR depending objective | Backup and DR solve different problems |
Notes and examples
Azure Backup components
| Component | Purpose | Exam cues |
|---|---|---|
| Recovery Services vault | Management container for backup/ASR items | Policies, protected items, jobs, alerts |
| MARS agent | Back up files/folders/system state from Windows Server | Common for on-prem Windows Server backup to Azure |
| Microsoft Azure Backup Server | Protect workloads and servers through a backup server model | App-aware workload protection scenarios |
| Backup policy | Schedule and retention | Match recovery need; do not invent retention from scenario |
| Recovery point | Point in time available for restore | Application-consistent vs crash-consistent may matter |
| Soft delete / immutability-style protections | Protect against accidental or malicious deletion where configured | Security and recovery controls are separate from backup schedule |
Azure Site Recovery components
| Component | Purpose | Exam cues |
|---|---|---|
| Replication policy | Frequency/retention/app consistency behavior | Drives RPO-related behavior |
| Mobility service / provider components | Replication agents/components depending source platform | Health must be monitored |
| Recovery plan | Ordered failover groups and automation steps | Multi-tier app failover |
| Test failover | Non-disruptive validation | Always preferred before real failover |
| Planned failover | Controlled failover when source is available | Minimizes data loss |
| Unplanned failover | Disaster scenario | Requires post-failover validation |
| Failback | Return workloads after primary site recovery | Must be planned and tested |
DR decision checklist
- Define the workload dependency map: identity, DNS, database, file shares, certificates, IP dependencies.
- Determine RPO/RTO from the scenario, then choose backup, replication, clustering, or ASR.
- Verify network design: Azure VNets, subnets, DNS, routing, VPN/ExpressRoute, NSGs, firewalls.
- Configure replication or backup policy.
- Run test failover or test restore.
- Document cutover order, validation steps, and rollback.
- Monitor jobs, agent health, replication health, and recovery point availability.
Azure Backup
Azure Backup is for protected recovery points and restore operations.
| Backup target | Typical approach |
|---|---|
| Azure VMs | Azure VM backup |
| On-premises files/folders/system state | Microsoft Azure Recovery Services agent or related backup architecture |
| Workloads at scale | Azure Backup with appropriate agents, vaults, and policies |
| System state recovery | Use supported backup method for Windows Server system state |
Know the workflow:
- Create or use a Recovery Services vault.
- Configure backup policy.
- Register/protect the workload.
- Run initial backup.
- Monitor jobs and alerts.
- Test restore procedures.
Common traps:
- Backups must be restorable; a configured backup policy is not enough.
- System state backup is different from full application-aware workload protection.
- Backup helps with corruption, deletion, ransomware recovery, and point-in-time restore; replication alone may not.
Azure Site Recovery
Azure Site Recovery focuses on workload replication and orchestrated failover.
| Requirement | ASR relevance |
|---|---|
| Replicate VMs to Azure | Strong fit |
| Test disaster recovery without disrupting production | Strong fit when test failover is supported/configured |
| Create recovery plans with ordered failover | Strong fit |
| Keep long-term historical restore points | Backup is usually the better concept |
| Protect individual files only | Backup or file-level solutions are usually better |
ASR decision clues:
- “Fail over workloads to Azure”
- “Orchestrate recovery”
- “Minimize downtime during site outage”
- “Run a test failover”
- “Replicate VMs”
Hyper-V Replica
Hyper-V Replica replicates VMs between Hyper-V hosts or clusters.
Use it when the scenario is specifically about Hyper-V-based replication and does not require broader Azure recovery orchestration.
Trap: Hyper-V Replica is not the same as failover clustering. Clustering handles high availability within the cluster; Replica handles VM replication for recovery.
Disaster recovery decision path
flowchart TD
A[What is the protection goal?] --> B[Restore deleted/corrupt data]
A --> C[Keep workload running after host failure]
A --> D[Fail over to another site or Azure]
A --> E[Replicate storage volumes]
B --> F[Use backup and tested restores]
C --> G[Use failover clustering / HA design]
D --> H[Use Azure Site Recovery or Hyper-V Replica]
E --> I[Use Storage Replica where appropriate]
Migration Reference
Migration service selection
| Source / target scenario | Prefer | Key reason | Common trap |
|---|---|---|---|
| Assess server estate for Azure readiness | Azure Migrate discovery and assessment | Inventory, sizing, dependency analysis | Discovery does not move workloads |
| Rehost VMware/Hyper-V/physical server to Azure VM | Azure Migrate server migration | Replication and cutover workflow | Lift-and-shift may still require app remediation |
| Move file server to newer Windows Server or Azure VM | Storage Migration Service | Inventories data, shares, ACLs, and supports cutover | Not continuous sync after migration |
| Move file data into Azure Files | Azure File Sync, AzCopy, Robocopy, or migration tooling | Depends on ongoing cache/sync vs one-time copy | Azure File Sync is not just a copy command |
| Keep branch file server cache with cloud namespace | Azure File Sync | Local cache plus Azure Files centralization | Plan sync groups and endpoint layout |
| Move AD DS to newer servers | Add new DCs, transfer FSMO roles, demote old DCs | Supported modernization path | Do not clone/restore DCs carelessly |
| Move IIS apps | Web Deploy, Azure Migrate/app assessment, App Service tools where applicable | Depends on rehost vs refactor | App dependencies may block simple move |
| Move databases | Database-specific migration tooling | Schema, compatibility, downtime requirements | File copy is not database migration |
Notes and examples
Storage Migration Service flow
| Phase | What happens | Validate |
|---|---|---|
| Prepare orchestrator | Install/administer Storage Migration Service | Network, firewall, permissions |
| Inventory source | Discover shares, files, security, local users/groups | Source access and complete inventory |
| Transfer data | Copy data to destination | ACLs, timestamps, share paths |
| Cut over | Destination assumes source name/IP where configured | Client access, DNS, application paths |
| Decommission | Remove or repurpose old server after validation | Backups and rollback window |
Azure Migrate flow
| Phase | Focus | Candidate reminders |
|---|---|---|
| Discover | Deploy appliance or agent-based discovery as required | Credentials, network reachability, inventory scope |
| Assess | Readiness, sizing, dependencies | Assessment assumptions affect recommendations |
| Remediate | Fix OS, disk, boot, app, network, identity issues | Do not migrate known-broken dependencies |
| Replicate | Start replication to Azure | Monitor replication health |
| Test migrate | Validate isolated or test environment | Avoid production DNS/IP conflicts |
| Cut over | Stop source changes and migrate | Plan downtime and rollback |
| Optimize | Rightsize, secure, back up, monitor | Migration is not complete until operations are configured |
Azure Migrate
Azure Migrate is commonly used to assess and migrate servers to Azure.
| Phase | What to know |
|---|---|
| Discovery | Inventory servers and dependencies |
| Assessment | Evaluate readiness, sizing, cost, and compatibility |
| Replication | Prepare migration by copying workload data |
| Test migration | Validate before production cutover |
| Cutover | Finalize migration with planned downtime as required |
- Assessment and migration are not the same step.
- Dependency mapping matters for multi-tier applications.
- Sizing should reflect observed utilization, not just allocated resources.
- Network, identity, DNS, and firewall dependencies can break an otherwise successful server migration.
- Test migration reduces risk; it does not replace application validation.
Storage Migration Service
Storage Migration Service is high-yield for file server migrations.
| It helps migrate… | Important details |
|---|---|
| Shares | Share names and paths must be planned |
| Files and folders | Permissions and ownership matter |
| Server identity | Cutover can preserve client access patterns |
| Legacy file servers | Useful when moving from older Windows Server file servers |
Decision rule: if the scenario says “migrate file servers while preserving shares, permissions, and server identity,” think Storage Migration Service.
Trap: copying files manually may lose permissions, share configuration, timestamps, or client access continuity.
Azure File Sync
Azure File Sync synchronizes on-premises Windows Server file shares with Azure Files.
| Feature | Meaning |
|---|---|
| Cloud endpoint | Azure file share |
| Server endpoint | Path on a registered Windows Server |
| Sync group | Relationship between cloud and server endpoints |
| Cloud tiering | Keeps frequently used files local and tiers cooler data to Azure |
| Registered server | On-premises server participating in sync |
Common mistakes:
- Treating Azure File Sync as backup. It synchronizes changes, including unwanted changes.
- Forgetting that users may still access local file servers while data synchronizes with Azure Files.
- Ignoring bandwidth, initial sync time, and namespace design.
Migration choice table
| Scenario language | Best concept to consider |
|---|---|
| “Assess on-premises servers before moving to Azure” | Azure Migrate assessment |
| “Move VMs to Azure with minimal guesswork about sizing” | Azure Migrate with assessment data |
| “Migrate a file server and keep shares/permissions” | Storage Migration Service |
| “Keep branch file access local while centralizing in Azure” | Azure File Sync |
| “Replicate VMs for disaster recovery” | Azure Site Recovery, not a migration-only tool |
| “Move application with databases and dependencies” | Dependency mapping, app validation, staged migration |
Monitoring, Logging, and Alerting
Monitoring component selection
| Need | Use | Notes |
|---|---|---|
| Guest OS event/performance collection | Azure Monitor Agent + DCR | DCR defines what to collect and where to send it |
| Query logs | Log Analytics workspace | KQL-based analysis |
| Visualize VM performance/dependencies | VM insights | Requires appropriate agent/configuration |
| Alert on log pattern | Azure Monitor log alert | Query returns condition over time |
| Alert on metric threshold | Azure Monitor metric alert | Lower-latency for platform metrics |
| Monitor backup jobs | Backup center / vault jobs and alerts | Check job status and protected item health |
| Monitor ASR replication | Recovery Services vault replication health | Look at agent and replication status |
| Monitor security posture | Defender for Cloud | Recommendations, alerts, secure score |
| Track update compliance | Azure Update Manager | Assessment and deployment results |
| Troubleshoot Azure network path | Network Watcher / Connection Monitor | Especially useful for Azure networking dependencies |
Notes and examples
KQL patterns to recognize
Table availability depends on the agent, DCR, workspace, and solution configuration.
Heartbeat
| where TimeGenerated > ago(1h)
| summarize LastHeartbeat=max(TimeGenerated) by Computer
| order by LastHeartbeat asc
Event
| where TimeGenerated > ago(24h)
| where EventLog == "System"
| where EventLevelName in ("Error", "Critical")
| summarize Count=count() by Computer, Source, EventID
| order by Count desc
Perf
| where TimeGenerated > ago(1h)
| where ObjectName == "LogicalDisk"
| where CounterName == "% Free Space"
| summarize LatestFreePercent=arg_max(TimeGenerated, CounterValue) by Computer, InstanceName
Update
| where TimeGenerated > ago(7d)
| summarize Updates=count() by Computer, Classification
Alert design checklist
| Check | Why it matters |
|---|---|
| Correct target scope | Alerts scoped too narrowly miss servers |
| Correct signal type | Metrics, logs, activity logs, and service health are different |
| Action group configured | Alert without notification or automation may be useless |
| Evaluation frequency/window | Too short causes noise; too long delays response |
| Suppression/maintenance plan | Avoid false positives during planned patching |
| Runbook or remediation path | Candidates should connect alerts to action |
Troubleshooting Cheat Sheet
First-pass hybrid troubleshooting workflow
- Confirm identity and authorization: Azure RBAC, local admin rights, domain membership, managed identity/service principal.
- Confirm DNS and time: name resolution, DC locator, Kerberos time requirements.
- Confirm network path: firewall, proxy, routing, TLS inspection, NSG, Windows Defender Firewall.
- Confirm agent health: Arc, AMA, MARS, ASR mobility/provider, Defender onboarding.
- Confirm configuration scope: policy assignment, DCR association, backup policy, update schedule, vault registration.
- Check logs: Event Viewer, agent logs, Azure activity logs, Log Analytics, service-specific job history.
- Test with a minimal path: one server, one rule, one workspace/vault, one known event.
Symptom-to-check table
| Symptom | Check first | Useful direction |
|---|---|---|
| Server not visible in Azure as Arc-enabled | Connected Machine agent, outbound connectivity, proxy, resource group/RBAC | Reconnect or re-onboard after fixing connectivity/identity |
| Arc server visible but no logs | AMA installed, DCR associated, workspace target, data source configured | Install/repair AMA and apply DCR |
| Update assessment missing | Azure Update Manager eligibility, Arc/VM status, agent health | Trigger assessment after agent/connectivity fix |
| Defender recommendation not appearing | Defender for Cloud plan, agent/extension, scope, policy | Confirm subscription/workspace/server onboarding |
| Azure Backup job failing | MARS/MABS/extension health, vault credentials, VSS writers, network | Check job error and local event logs |
| ASR replication unhealthy | Mobility service/provider, process components, replication policy, network | Re-sync or repair agent after root cause |
| Domain join fails | DNS points to AD DS DNS, domain reachability, credentials, time | Test name resolution and DC locator |
| Kerberos/auth failures | Time skew, SPNs, duplicate names, secure channel | Use w32tm, setspn, nltest |
| Cluster resource fails over unexpectedly | Cluster logs, resource dependencies, storage/network health, witness | Validate cluster and inspect event logs |
| File migration permissions wrong | ACL translation, local users/groups, domain trust, SID history/mapping | Re-run validation before cutover |
| Slow Azure VM domain logon | AD Sites and Services subnets, DNS, DC placement | Add correct subnets and local DC/DNS path |
Notes and examples
Command reference
## AD DS health
dcdiag /v
repadmin /replsummary
nltest /dsgetdc:contoso.com
w32tm /query /status
Test-ComputerSecureChannel
## Network tests
Test-NetConnection dc01.contoso.com -Port 53
Test-NetConnection server01.contoso.com -Port 5985
Resolve-DnsName _ldap._tcp.dc._msdcs.contoso.com
## Azure Arc agent
azcmagent show
azcmagent check
azcmagent logs
## Cluster checks
Get-ClusterNode
Get-ClusterGroup
Get-ClusterResource
Get-ClusterQuorum
## Backup / VSS checks
vssadmin list writers
wbadmin get status
Azure Monitor and Log Analytics
Azure Monitor collects and analyzes telemetry. Log Analytics is commonly used for querying collected logs.
| Need | Concept |
|---|---|
| Query logs across servers | Log Analytics workspace |
| Collect Windows events | Agent/data collection configuration |
| Alert on conditions | Azure Monitor alerts |
| Visualize trends | Workbooks, metrics, dashboards |
| Investigate security posture | Defender for Cloud plus logs/recommendations |
| Manage hybrid server visibility | Azure Arc plus monitoring configuration |
KQL basics to recognize:
| Pattern | Meaning |
|---|---|
where | Filter rows |
summarize | Aggregate results |
count | Count records |
project | Select columns |
order by | Sort results |
| Time filters | Narrow results to a relevant investigation window |
Trap: Log Analytics only shows data that has been collected and sent. If a server is not connected, configured, or authorized, queries will not magically return its logs.
Windows Server troubleshooting checklist
Use this order when a scenario gives symptoms but not the cause:
- Scope — one user, one server, one subnet, one site, or all systems?
- Recent change — patch, GPO, certificate, DNS, firewall, route, storage, identity?
- Identity — authentication, authorization, Kerberos, SPN, time sync?
- Name resolution — DNS records, suffixes, conditional forwarders, stale records?
- Network path — firewall, routing, NSG if Azure, VPN/ExpressRoute, ports?
- Service health — service status, event logs, dependencies?
- Performance — CPU, memory, disk latency, queue length, network throughput?
- Storage — free space, permissions, locks, replication status?
- Cluster state — node status, quorum, role ownership, CSV health?
- Logs and metrics — correlate time of failure with events.
Common symptom-to-cause map
| Symptom | High-yield checks |
|---|---|
| Users cannot access file share | DNS, SMB port/firewall, share permissions, NTFS permissions, server service |
| Admin cannot connect remotely | WinRM/RDP enabled, firewall, local policy, credentials, network path |
| Kerberos authentication fails | Time sync, SPN, DNS, domain controller reachability |
| Cluster role will not fail over | Dependencies, storage, network name, quorum, node health |
| VM migration fails | CPU compatibility, network, storage, permissions, cluster configuration |
| Slow file access | Disk latency, network latency, SMB settings, antivirus scanning, tiering state |
| Backup job fails | Agent health, vault registration, credentials, storage, VSS writers |
| ASR replication unhealthy | Connectivity, agent/provider health, storage churn, credentials, replication policy |
| Azure Arc server offline | Agent service, outbound connectivity, proxy, identity, permissions |
| Certificate-based service fails | Expiration, trust chain, subject/SAN, private key, binding |
High-Yield Exam Distinctions
| Distinction | Remember |
|---|---|
| Azure Backup vs Azure Site Recovery | Backup restores recovery points; ASR orchestrates workload failover |
| Failover clustering vs ASR | Clustering is HA; ASR is DR/failover orchestration |
| Storage Replica vs DFS Replication | Storage Replica is block-level volume replication; DFSR is file-level replication |
| Azure Arc vs Azure Migrate | Arc manages existing machines; Azure Migrate moves/assesses workloads |
| Azure Monitor Agent vs Log Analytics workspace | Agent collects; workspace stores/query logs; DCR defines collection |
| Defender Antivirus vs Defender for Cloud | Antivirus protects endpoint; Defender for Cloud assesses and protects cloud/hybrid posture |
| NSG vs Windows Defender Firewall | NSG filters Azure network traffic; host firewall filters inside the OS |
| Microsoft Entra ID vs AD DS | Entra ID is cloud identity; AD DS provides domain services, Kerberos, GPO, LDAP |
| Cloud witness vs backup | Witness participates in quorum; it stores no protected workload data |
| Test failover vs planned failover | Test failover validates DR without production cutover; planned failover is controlled production move |
| Azure File Sync vs Storage Migration Service | File Sync supports ongoing sync/cache; SMS is migration/cutover focused |
| WAC vs Azure portal | WAC manages Windows Server directly; Azure portal manages Azure resources and Arc projections |
Final Review Checklist
Before sitting for AZ-802, make sure you can quickly answer:
- Which service manages non-Azure Windows Servers through Azure Resource Manager?
- Which component controls Azure Monitor Agent data collection?
- When would you choose Azure Backup instead of Azure Site Recovery?
- How do quorum and witness settings prevent split-brain in a cluster?
- Which tool migrates file servers while preserving shares and ACLs?
- How do you validate a cluster before creating it?
- What breaks domain join for Azure VMs most often?
- How do you troubleshoot Arc, AMA, backup, and ASR agent health?
- How do AD DS, Microsoft Entra ID, and Microsoft Entra Domain Services differ?
- What should be tested before a real DR failover?
Next step: convert the decision tables into scenario flashcards, then complete timed AZ-802 practice questions that force you to choose the correct Microsoft service, agent, policy, or recovery pattern from a short business requirement.
Notes and examples
Final review checklist
Before moving into original practice questions, make sure you can explain:
- When to use Azure Backup versus Azure Site Recovery.
- When to use failover clustering versus Hyper-V Replica.
- How quorum and witnesses affect cluster availability.
- Why Storage Replica is not backup.
- How Azure Arc changes hybrid server management.
- How Azure Monitor, Log Analytics, and agents fit together.
- How Storage Migration Service differs from Azure File Sync.
- How Azure Migrate assessment differs from migration execution.
- How to troubleshoot DNS, Kerberos, firewall, and certificate issues.
- How to apply least privilege to Windows Server administration.
- How to patch clustered workloads safely.
- How to read scenario clues around downtime, RPO/RTO, identity, and connectivity.
High-yield AZ-802 mental model
Think in five connected workstreams:
| Workstream | What the exam often tests | Fast decision point |
|---|---|---|
| Secure Windows Server | Least privilege, identity protection, hardening, update posture, Defender integrations | Is the risk identity, endpoint, network, data, or admin access? |
| Implement high availability | Failover clustering, load balancing, Storage Spaces Direct, Cluster-Aware Updating | Is the goal local availability or regional/site recovery? |
| Disaster recovery and backup | Azure Backup, Azure Site Recovery, Hyper-V Replica, Storage Replica | Is the goal restore, failover, replication, or rollback? |
| Migrate servers and workloads | Azure Migrate, Storage Migration Service, Windows Admin Center, data/app migration | Are you moving compute, storage, identity, or application dependencies? |
| Monitor and troubleshoot | Azure Monitor, Log Analytics, Azure Arc, event logs, performance counters | Is the problem resource health, OS behavior, network, identity, or application performance? |
Hybrid management with Azure Arc and Windows Admin Center
Azure Arc-enabled servers
Azure Arc is central to hybrid operations. It lets you project non-Azure Windows Servers into Azure for management.
| Capability | What it enables |
|---|---|
| Inventory and governance | See hybrid servers as Azure resources |
| Policy and compliance | Apply Azure Policy where supported |
| Monitoring integration | Send logs and metrics to Azure Monitor / Log Analytics |
| Security posture | Surface recommendations through Microsoft security tooling |
| Extension management | Install supported agents/extensions from Azure |
Notes and examples
Decision rule: choose Azure Arc when the server remains on-premises or outside Azure but needs Azure-based management, governance, monitoring, or security visibility.
Trap: Azure Arc does not automatically convert a server into an Azure VM and does not eliminate the need for network connectivity, permissions, or agents.
Windows Admin Center
Windows Admin Center is useful for managing Windows Server roles, failover clusters, Hyper-V, updates, certificates, storage, and Azure integrations.
| Use Windows Admin Center for… | Do not confuse it with… |
|---|---|
| Browser-based server administration | A replacement for all enterprise monitoring |
| Managing Hyper-V and clusters | Azure Site Recovery orchestration by itself |
| Azure hybrid service onboarding | The Azure control plane itself |
| Certificate, event, service, and role management | A substitute for security governance |
Exam clue: if the question asks for a practical management tool for on-premises Windows Server with optional Azure integrations, Windows Admin Center is often relevant.
Networking and hybrid connectivity review
AZ-802 may embed networking details inside migration, backup, monitoring, and hybrid management scenarios.
| Area | Know this |
|---|---|
| DNS | AD DS, Kerberos, file access, and app connectivity depend heavily on correct name resolution |
| Firewall rules | Prefer specific allowed ports over disabling firewalls |
| VPN/ExpressRoute | Connectivity choice affects latency, routing, resilience, and private access |
| Private endpoints | Used to access supported Azure services privately where configured |
| Proxies | Hybrid agents often require outbound connectivity and proxy awareness |
| Time sync | Authentication and clustering can fail when time is inconsistent |
| Routing | Hybrid failures are often route table, gateway, or asymmetric routing issues |
Common trap: a server can appear “healthy” locally while Azure management fails because outbound connectivity, proxy configuration, or required identity permissions are missing.
Update and patch management
Patch questions often test service continuity, not just “install updates.”
| Environment | Review focus |
|---|---|
| Standalone servers | Maintenance windows, restart planning, rollback approach |
| Clusters | Drain roles, patch node, reboot, resume, repeat |
| Hybrid servers | Inventory, compliance visibility, Azure management integration |
| Security-sensitive systems | Prioritization, testing, emergency patch process |
| Domain controllers | Redundancy, replication health, staged patching |
Cluster patching rule: never think of a cluster as one server. Patch one node at a time, maintain quorum, drain workloads, and verify role health after each node.
Role-specific quick hits
File services
| Topic | Review point |
|---|---|
| NTFS vs share permissions | Effective access is constrained by both |
| Access-based enumeration | Hides folders users cannot access |
| FSRM | Quotas, file screens, classification/reporting |
| DFS Namespace | Logical namespace for shares |
| DFS Replication | File replication, not database replication |
| Azure File Sync | Hybrid file sync with Azure Files |
Notes and examples
Trap: “User cannot access a share” may be a permissions issue, a name resolution issue, a firewall issue, or a server availability issue. Do not jump directly to NTFS permissions without reading the symptom.
Hyper-V
| Topic | Review point |
|---|---|
| Checkpoints | Useful for some rollback scenarios but not a backup replacement |
| Live migration | Moves running VMs between hosts when configured |
| Replica | Replicates VMs for recovery |
| Shielded VMs | Protect VMs from compromised fabric administrators in supported environments |
| Virtual switches | External, internal, private connectivity models |
| Integration services | Affect guest operations and management |
Trap: checkpoints can create operational risk if left unmanaged, especially on production workloads.
Containers and application workloads
If containers appear, focus on the operational distinction:
| Concept | Review point |
|---|---|
| Windows containers | Process-isolated or Hyper-V-isolated Windows workloads |
| Image | Packaged application filesystem and configuration |
| Registry | Stores container images |
| Host compatibility | Windows container compatibility depends on host and image requirements |
| Orchestration | May involve broader platform choices outside basic server administration |
Do not over-focus on developer details unless the scenario specifically asks about container hosting, isolation, or compatibility.
Exam-style decision rules
Use these fast rules during practice:
- If the problem is visibility/governance for non-Azure servers, consider Azure Arc.
- If the problem is server migration to Azure, consider Azure Migrate.
- If the problem is file server migration with permissions and shares, consider Storage Migration Service.
- If the problem is branch/local file caching with Azure Files, consider Azure File Sync.
- If the problem is point-in-time recovery, consider Azure Backup.
- If the problem is orchestrated failover, consider Azure Site Recovery.
- If the problem is local workload availability, consider failover clustering.
- If the problem is cluster patching, consider Cluster-Aware Updating.
- If the problem is block-level volume replication, consider Storage Replica.
- If the problem is narrow delegated administration, consider Just Enough Administration.
- If the problem is unique local admin passwords, consider LAPS-style management.
- If the problem is centralized logs and queries, consider Azure Monitor and Log Analytics.