AZ-900 Azure Fundamentals
Compressed Course
AZ-900 Microsoft Azure Fundamentals โ Compressed Exam-Preparation Course
Alignment: AZ-900 skills measured as of January 14, 2026
Purpose: A fast but complete revision course built from the analyzed 1,100-question AZ-900 practice bank and organized around the current Microsoft blueprint.
Best use: Learn the decision rules first, then use practice questions to test whether you can select the correct Azure concept or service from plausible distractors.
1. Exam Overview
What the exam is testing
AZ-900 tests whether you can recognize foundational cloud concepts and choose the most appropriate Azure concept, service, or management tool for a short scenario.
You are not expected to design complex enterprise systems or configure services from memory. You are expected to answer questions such as:
- Is this public, private, or hybrid cloud?
- Is the best fit IaaS, PaaS, SaaS, or serverless?
- Does the scenario need a virtual machine, a container, a web app, or a function?
- Does the network requirement call for VNet peering, VPN Gateway, ExpressRoute, a public endpoint, or a private endpoint?
- Is the question about governance, cost, monitoring, security, or deployment?
- Which Azure tool solves the stated problem most directly?
The official passing score is 700 or greater. Treat every question as a decision problem: identify the requirement, eliminate unrelated service categories, and select the option that most directly solves the problem.
How to think like the exam
Most AZ-900 questions use a simple pattern:
- A company has a requirement.
- Several answers are real Azure concepts or services.
- Only one answer matches the primary requirement.
- The strongest distractor often solves a nearby but different problem.
Use this mental sequence:
- Identify the category: cloud concept, architecture, compute, network, storage, identity, cost, governance, deployment, or monitoring.
- Highlight the decisive words: private, over the public internet, without managing servers, prevent deletion, estimate cost before deployment, service incident, or application telemetry.
- Eliminate services from the wrong category.
- Choose the simplest Azure option that directly satisfies the requirement.
- Do not add complexity unless the scenario requires it.
How to use this course
Use the guide in four passes:
- Pass 1: Learn the maps and comparison tables.
- Pass 2: Practice service selection by covering the answer column and choosing from the scenario alone.
- Pass 3: Review the traps and explain why the best wrong answer fails.
- Pass 4: Use the rapid-review section and exam-day checklist.
2. Exam Domains
Domain list and priorities
| Official domain | Official weight | Study priority | Why it matters |
|---|---|---|---|
| Describe cloud concepts | 25โ30% | High | Establishes the vocabulary used in every other domain |
| Describe Azure architecture and services | 35โ40% | Very high | Largest domain; contains the most service-selection questions |
| Describe Azure management and governance | 30โ35% | Very high | Frequently tests tools that sound similar but solve different problems |
Priority notes
Spend the most time on:
- Azure architecture, compute, networking, storage, and identity.
- Governance, cost management, deployment tools, and monitoring.
- Cloud models, service models, shared responsibility, and cloud benefits.
What matters most
The exam repeatedly rewards accurate mapping:
| Requirement signal | Likely answer |
|---|---|
| Control the guest operating system | Azure Virtual Machines / IaaS |
| Host a web application without patching the OS | Azure App Service / PaaS |
| Run event-driven code without managing servers | Azure Functions / serverless |
| Encrypted connection over the public internet | Azure VPN Gateway |
| Private connectivity that avoids the public internet | ExpressRoute |
| Private IP access to an Azure service | Private endpoint |
| Apply standards across resources | Azure Policy |
| Prevent accidental deletion | Resource lock |
| Estimate cost before deployment | Azure pricing calculator |
| Analyze current spending and set budgets | Microsoft Cost Management |
| Platform incident or planned maintenance | Azure Service Health |
| Telemetry, logs, metrics, alerts | Azure Monitor |
| Application performance and failures | Application Insights |
| Personalized optimization recommendations | Azure Advisor |
3. Start-to-Finish Study Path
Foundation
Learn these first:
- Public, private, and hybrid cloud.
- Consumption-based pricing, CapEx, and OpEx.
- Shared responsibility.
- High availability, scalability, elasticity, reliability, predictability, governance, and manageability.
- IaaS, PaaS, SaaS, and serverless.
Checkpoint: you should be able to explain why a VM is usually IaaS, why App Service is PaaS, and why Microsoft 365 is SaaS.
Intermediate
Move to Azure architecture and core services:
- Regions, datacenters, availability zones, region pairs, and sovereign regions.
- Management groups, subscriptions, resource groups, and resources.
- Virtual machines, scale sets, availability sets, Azure Virtual Desktop, containers, Functions, and web apps.
- VNets, subnets, peering, Azure DNS, VPN Gateway, ExpressRoute, public endpoints, and private endpoints.
- Blob, Files, Queue, Table, managed disks, storage tiers, and redundancy.
- Microsoft Entra ID, Entra Domain Services, SSO, MFA, passwordless, external identities, Conditional Access, RBAC, Zero Trust, defense in depth, and Defender for Cloud.
Checkpoint: for each scenario, identify the service category before selecting a product.
Advanced
Focus on confusing pairs and multi-constraint questions:
- Availability zone versus availability set versus region pair.
- VPN Gateway versus ExpressRoute.
- Public endpoint versus private endpoint.
- Blob versus Files versus Queue versus Table versus managed disks.
- Microsoft Entra ID versus Entra Domain Services.
- SSO versus MFA versus passwordless versus Conditional Access.
- Azure Policy versus resource locks versus tags versus RBAC.
- Advisor versus Service Health versus Monitor versus Log Analytics versus Application Insights.
- Pricing calculator versus Cost Management.
- Portal versus Cloud Shell versus CLI versus PowerShell versus ARM templates.
Checkpoint: explain why the most tempting distractor is wrong.
Final review
Use the last-day method:
- Read the quick memory rules.
- Review the service-selection tables.
- Rehearse the architecture hierarchy.
- Revisit storage redundancy and monitoring tools.
- Answer a mixed set of questions without notes.
- Review only the concepts you missed.
4. Core Concepts by Domain
Domain 1 โ Describe Cloud Concepts
4.1 Cloud computing
Cloud computing delivers computing services over the internet or through a cloud environment. Instead of buying and operating every server yourself, you use resources as needed.
Cloud models
| Model | Meaning | Best fit | Common trap |
|---|---|---|---|
| Public cloud | Provider-owned cloud resources shared across customers with logical isolation | Fast deployment, broad cloud services, consumption-based usage | โPublicโ does not mean your data is automatically public |
| Private cloud | Cloud environment dedicated to one organization | Dedicated control or specialized internal requirements | Private cloud is not the same as a single private endpoint |
| Hybrid cloud | Combines on-premises or private-cloud resources with public-cloud services | Gradual migration, legacy dependencies, mixed environments | If any required workload remains on-premises while Azure is also used, consider hybrid |
Consumption-based model
You use resources when needed and generally pay according to usage. The key advantage is avoiding unnecessary upfront hardware purchases.
| Term | Meaning | Exam clue |
|---|---|---|
| Capital expenditure (CapEx) | Upfront purchase of assets such as physical servers | โBuy hardware before deploymentโ |
| Operational expenditure (OpEx) | Ongoing spending such as a variable cloud bill | โMonthly bill changes with usageโ |
| Consumption-based pricing | Charges align with resource usage | โStop paying after deleting or stopping resourcesโ |
Shared responsibility model
Cloud adoption changes who manages each layer.
| Layer | On-premises | IaaS | PaaS | SaaS |
|---|---|---|---|---|
| Physical datacenter | Customer | Microsoft | Microsoft | Microsoft |
| Physical network and hosts | Customer | Microsoft | Microsoft | Microsoft |
| Guest OS | Customer | Customer | Microsoft | Microsoft |
| Runtime and middleware | Customer | Customer | Microsoft | Microsoft |
| Application | Customer | Customer | Customer | Mostly Microsoft |
| Data, identities, and access decisions | Customer | Customer | Customer | Customer |
Rule: the more managed the service, the less infrastructure the customer manages. The customer still remains responsible for data governance, identity choices, and access decisions.
Serverless
Serverless computing lets you run code without managing servers. The cloud provider handles infrastructure management and scaling details.
Exam clue: event-driven code, short execution, trigger-based processing, no server management โ Azure Functions.
4.2 Benefits of cloud services
| Benefit | Meaning | Scenario clue | Do not confuse with |
|---|---|---|---|
| High availability | Keep a service accessible when failures occur | โRemain available if a component failsโ | Scalability |
| Scalability | Adjust capacity to handle demand | โIncrease capacityโ | High availability |
| Horizontal scaling | Add or remove instances | โAdd more VMs behind a load balancerโ | Vertical scaling |
| Vertical scaling | Increase or decrease capacity of one resource | โIncrease CPU and memory on a VMโ | Horizontal scaling |
| Elasticity | Scale dynamically as demand rises and falls | โSeasonal spike; add and remove resources automaticallyโ | Static scaling |
| Reliability | Recover from failures and continue operating | โReduce effect of failures; recoverโ | Cost optimization |
| Predictability | Improve confidence in cost or performance planning | โForecast expected usage or costโ | Guaranteed zero variation |
| Security | Use cloud controls and provider capabilities to protect resources | โProtect workloads and identitiesโ | Governance |
| Governance | Enforce standards and organizational rules | โApproved regions and required tagsโ | Monitoring |
| Manageability | Deploy and operate resources using portals, CLI, APIs, templates, and automation | โProvision through automationโ | Manual administration only |
4.3 Cloud service types
| Service model | Customer manages | Provider manages | Best scenario |
|---|---|---|---|
| IaaS | Guest OS, applications, data, configuration | Physical datacenter, hosts, underlying infrastructure | Lift-and-shift, custom OS control |
| PaaS | Application code and data | Infrastructure, OS, runtime, middleware | Managed application hosting |
| SaaS | Users, data usage, access decisions | Complete application and supporting platform | Ready-to-use business application |
Fast comparison
| Requirement | Choose |
|---|---|
| Maximum guest OS control | IaaS |
| Deploy code without patching the OS | PaaS |
| Use a complete provider-managed application | SaaS |
| Run triggered code without server management | Serverless |
Domain 1 traps
- Do not confuse elasticity with scalability. Elasticity emphasizes dynamic adjustment as demand changes.
- Do not confuse high availability with horizontal scaling. Scaling may support availability, but the concepts are not identical.
- Do not say Microsoft patches the guest OS of an IaaS VM. The customer manages the VM guest OS.
- Do not choose SaaS for a lift-and-shift migration. SaaS replaces the application with a complete managed application.
- Do not choose PaaS when the requirement explicitly demands guest OS control.
Domain 2 โ Describe Azure Architecture and Services
4.4 Core Azure architecture
Physical and geographic concepts
| Concept | Meaning | Key exam clue |
|---|---|---|
| Azure datacenter | Physical facility containing servers, networking, power, and cooling | โPhysical facilityโ |
| Azure region | Geographic area containing one or more datacenters connected through a low-latency network | โDeploy in a geographic areaโ |
| Availability zone | Physically separate location within a supported region with independent power, cooling, and networking | โProtect from datacenter-level failure in the same regionโ |
| Region pair | Relationship between two Azure regions used for certain replication and recovery considerations | โSecondary regionโ or regional recovery relationship |
| Sovereign region | Specialized Azure environment for particular government or legal requirements | โGovernment, legal, or isolated environmentโ |
Management hierarchy
Memorize this order:
Management group โ Subscription โ Resource group โ Resource
| Scope | Purpose | Exam clue |
|---|---|---|
| Management group | Organize and govern multiple subscriptions | โApply policy across subscriptionsโ |
| Subscription | Billing and access-control boundary | โSeparate billing for development and productionโ |
| Resource group | Logical container for resources managed together | โGroup resources for one application lifecycleโ |
| Resource | Individual Azure item, such as a VM, VNet, or storage account | โThe deployed service itselfโ |
Hierarchy traps
- A resource group is logical, not physical.
- A resource belongs to one resource group at a time.
- Resources inside one resource group can often be located in different regions.
- A subscription is not an availability or resiliency feature.
- A management group does not contain VMs directly; it organizes subscriptions.
4.5 Compute services
Compute selection table
| Requirement | Recommended service | Why | Common wrong answer |
|---|---|---|---|
| Full guest OS control | Azure Virtual Machines | VM provides OS-level control | Azure Functions |
| Identical load-balanced VMs that can scale | Virtual Machine Scale Sets | Manages a scalable VM group | Availability set |
| Spread VMs across fault and update domains | Availability set | Helps reduce impact of host maintenance or faults | Availability zone |
| Virtualized desktops and apps for remote users | Azure Virtual Desktop | Delivers desktop and app virtualization | Azure Functions |
| Portable package with code and dependencies | Containers | Consistent deployment unit | Resource group |
| Event-driven execution without server management | Azure Functions | Serverless compute | Azure Virtual Machines |
| Managed hosting for web applications | Azure App Service Web Apps | PaaS-style web hosting | VM unless OS control is required |
Azure VM supporting resources
A VM commonly needs:
- A network interface (NIC).
- A virtual network and subnet.
- Storage for the operating-system disk.
- Optional data disks.
- Optional public IP or private connectivity depending on the design.
Exam rule: a NIC connects a VM to a VNet. Managed disks provide persistent block storage for VM OS and data disks.
Availability set versus availability zone versus scale set
| Feature | Primary purpose |
|---|---|
| Availability set | Distribute VMs across fault and update domains |
| Availability zone | Separate physical location inside a region |
| Virtual Machine Scale Set | Deploy and scale a group of similar VMs |
Compute traps
- Choose Functions for event-driven serverless code, not for OS administration.
- Choose App Service for managed web hosting, not if the scenario explicitly requires guest OS control.
- Choose containers when portability and packaged dependencies matter.
- Choose VM Scale Sets when the question emphasizes a scalable group of similar VMs.
- Choose Azure Virtual Desktop for remote user desktops, not ordinary VM autoscaling.
4.6 Networking services
Core networking concepts
| Service or concept | Use it for | Key clue |
|---|---|---|
| Azure virtual network (VNet) | Private network address space for Azure resources | โPrivate network in Azureโ |
| Subnet | Segment a VNet into smaller ranges | โDivide a VNetโ |
| VNet peering | Private connectivity between VNets over the Microsoft backbone | โConnect two VNets privatelyโ |
| Azure DNS | Host DNS zones and resolve names | โName resolutionโ |
| Azure VPN Gateway | Encrypted connection over the public internet | โEncrypted tunnel over internetโ |
| ExpressRoute | Private connectivity from on-premises to Microsoft cloud services | โAvoid the public internetโ |
| Public endpoint | Access through a public address | โPublicly reachable service endpointโ |
| Private endpoint | Private IP address in a VNet for a supported Azure service | โReach storage or another service through a private IPโ |
VPN Gateway versus ExpressRoute
| Question clue | Choose |
|---|---|
| Encrypted connection using the public internet | VPN Gateway |
| Private circuit that does not send traffic over the public internet | ExpressRoute |
| Connect two Azure VNets privately | VNet peering |
| Give an Azure service a private IP inside a VNet | Private endpoint |
Networking traps
- ExpressRoute is not the answer merely because a connection is secure. It is the answer when private connectivity avoiding the public internet is required.
- VPN Gateway uses encrypted traffic over the public internet.
- VNet peering connects VNets; it is not a DNS service.
- A private endpoint is not the same as a subnet. It places a private IP for a supported service in your VNet.
- Azure DNS resolves names. It does not establish network connectivity.
4.7 Storage services
Storage service selection
| Service | Purpose | Example |
|---|---|---|
| Azure Blob Storage | Unstructured object data | Images, videos, backups, documents |
| Azure Files | Managed file shares | SMB file share for users or applications |
| Azure Queue Storage | Asynchronous messages | Decouple application components |
| Azure Table Storage | NoSQL key-attribute data | Large structured non-relational entity data |
| Azure managed disks | Persistent block storage for Azure VMs | OS disks and data disks |
Storage tiers
| Tier | Best for | Access pattern |
|---|---|---|
| Hot | Frequently accessed data | Regular reads and writes |
| Cool | Infrequently accessed but still online data | Occasional reads |
| Archive | Rarely accessed long-term data | Retrieval delay is acceptable |
Storage redundancy
| Option | Replication pattern | Choose when |
|---|---|---|
| LRS | Multiple copies in one datacenter in the primary region | Lowest-cost local redundancy |
| ZRS | Copies across availability zones in the primary region | Protect against zone-level failure |
| GRS | Copies in the primary region and a paired secondary region | Geographic replication |
| GZRS | Zone redundancy in the primary region plus replication to a secondary region | Combine zone and geographic resiliency |
Storage account options
A storage account provides a unique namespace for Azure Storage data. Important exam ideas include:
- Standard versus Premium performance.
- Storage services available through the account.
- Redundancy choice.
- Data-access pattern.
- Access tier for Blob Storage where relevant.
Data movement and migration
| Requirement | Tool |
|---|---|
| Command-line copying to or from Azure Storage | AzCopy |
| Graphical desktop management of Azure Storage data | Azure Storage Explorer |
| Cache an Azure file share on Windows Server and synchronize changes | Azure File Sync |
| Assess and plan migration of on-premises servers | Azure Migrate |
| Transfer a very large dataset offline because the network is too slow | Azure Data Box |
Storage traps
- Blob is object storage; Azure Files is a managed file share.
- Queue Storage stores messages, not files.
- Table Storage stores non-relational entities, not VM disks.
- Managed disks are for VM block storage.
- Archive is not for immediate access; retrieval takes time.
- ZRS stays in the primary region across zones.
- GRS adds a secondary geographic region.
- GZRS combines zones in the primary region with geographic replication.
- Azure Data Box transfers large offline datasets. It is not a monitoring, governance, or identity tool.
- Azure Migrate assesses and plans migrations. It is broader than offline data transfer.
4.8 Identity, access, and security
Identity and directory services
| Service | Purpose | Exam clue |
|---|---|---|
| Microsoft Entra ID | Cloud identity and access management | โUsers, applications, and access in the cloudโ |
| Microsoft Entra Domain Services | Managed domain capabilities such as domain join, LDAP, Kerberos, and NTLM without managing domain controllers | โLegacy domain features without deploying domain controllersโ |
| External identities | Collaboration with users outside the organization | โPartners use their own identitiesโ |
Authentication and access controls
| Capability | Purpose | Exam clue |
|---|---|---|
| Single sign-on (SSO) | Sign in once and access multiple authorized apps | โAvoid repeated sign-insโ |
| Multifactor authentication (MFA) | Require an additional verification factor | โBeyond a passwordโ |
| Passwordless authentication | Authenticate without entering a password | โBiometrics or security keyโ |
| Conditional Access | Apply access rules based on signals such as location, user, device, risk, or app | โRequire MFA from unfamiliar locationsโ |
| Azure RBAC | Grant authorized actions at a defined Azure scope | โAllow support team to restart VMs but not manage storageโ |
Security concepts
| Concept | Meaning |
|---|---|
| Zero Trust | Verify explicitly, use least privilege, and assume breach |
| Defense in depth | Apply multiple security layers: physical, identity, perimeter, network, compute, application, and data |
| Microsoft Defender for Cloud | Improve security posture and protect Azure and hybrid workloads |
Identity and security traps
- Entra ID is cloud identity and access management.
- Entra Domain Services provides managed legacy domain capabilities without customer-managed domain controllers.
- SSO improves sign-in experience. It does not automatically add a second factor.
- MFA adds verification factors. It is not the same as passwordless.
- Conditional Access decides when access controls apply.
- RBAC decides what an identity can do at a scope.
- Defender for Cloud improves workload security posture. It is not the same as Entra ID.
Domain 3 โ Describe Azure Management and Governance
4.9 Cost management
Cost tools
| Requirement | Tool | Why |
|---|---|---|
| Estimate expected price before deployment | Azure pricing calculator | Models planned services and configurations |
| Analyze spending, create budgets, and review trends | Microsoft Cost Management | Manages actual and forecasted spending |
| Identify owner, department, or environment | Tags | Adds metadata for organization and reporting |
| Find cost-optimization recommendations | Azure Advisor | Suggests improvements such as reducing waste |
Cost factors
Azure cost can vary with:
- Service type.
- Resource size or tier.
- Usage duration.
- Region.
- Data transfer patterns.
- Storage redundancy and access tier.
- Number of deployed resources.
Cost traps
- Tags help organize and report on resources. Tags do not automatically reduce charges.
- A budget can notify you when spending approaches a threshold. It does not automatically fix every cost issue.
- The pricing calculator estimates a future design. Cost Management analyzes spending in an Azure environment.
- Azure Advisor recommends improvements. It is not the main budgeting tool.
4.10 Governance and compliance
Governance tools
| Requirement | Tool | Why |
|---|---|---|
| Require approved regions or tags | Azure Policy | Audit or enforce standards |
| Prevent accidental deletion but allow changes | CanNotDelete lock | Blocks deletion |
| Prevent changes and deletion | ReadOnly lock | More restrictive lock |
| Discover and govern data across an environment | Microsoft Purview | Data governance capability |
| Apply governance across subscriptions | Management group plus Azure Policy | Parent scope for child subscriptions |
Azure Policy versus resource lock versus RBAC versus tags
| Tool | Main question it answers |
|---|---|
| Azure Policy | Is this resource configuration allowed or compliant? |
| Resource lock | Can this resource be deleted or modified? |
| Azure RBAC | What actions can this identity perform at this scope? |
| Tags | How should this resource be classified or reported? |
Governance traps
- Azure Policy does not replace RBAC.
- RBAC controls authorized actions; Policy evaluates allowed configurations.
- A CanNotDelete lock still allows authorized changes.
- A ReadOnly lock blocks changes and deletion.
- Tags are metadata, not security boundaries.
- Microsoft Purview is about data governance, not VM monitoring.
4.11 Management and deployment tools
Management interfaces
| Tool | Best use |
|---|---|
| Azure portal | Browser-based graphical resource management |
| Azure Cloud Shell | Authenticated browser shell with Bash or PowerShell experiences |
| Azure CLI | Cross-platform az commands |
| Azure PowerShell | PowerShell cmdlets for Azure administration |
| Azure Arc | Extend Azure management to servers and Kubernetes resources outside Azure |
| Infrastructure as code (IaC) | Define repeatable deployments in version-controlled files |
| Azure Resource Manager (ARM) | Azure deployment and management layer |
| ARM template | Declarative JSON file for repeatable Azure deployments |
Deployment decision rules
| Scenario | Choose |
|---|---|
| Beginner wants a graphical browser interface | Azure portal |
| Administrator wants a browser shell | Cloud Shell |
Linux-oriented automation with az commands |
Azure CLI |
| PowerShell-based automation | Azure PowerShell |
| Manage servers outside Azure through Azure capabilities | Azure Arc |
| Repeatable version-controlled infrastructure deployment | IaC |
| Declarative JSON deployment | ARM template |
Deployment traps
- Cloud Shell is a browser-accessible shell; it is not the same as the portal GUI.
- CLI and PowerShell are both management tools. Select based on the command style described.
- ARM is the management layer; an ARM template is a declarative deployment file.
- Azure Arc extends management beyond Azure. It is not a data-transfer appliance.
4.12 Monitoring tools
Monitoring selection
| Requirement | Tool |
|---|---|
| Personalized recommendations for cost, reliability, security, performance, or operations | Azure Advisor |
| Azure platform incident, planned maintenance, or health advisory | Azure Service Health |
| Collect, analyze, and act on metrics and logs | Azure Monitor |
| Query and analyze collected log data | Log Analytics |
| Notify when a threshold is crossed | Azure Monitor alerts |
| Monitor application performance, request rates, failures, and dependencies | Application Insights |
The monitoring map
Think of the tools as a sequence:
- Advisor: โWhat should I improve?โ
- Service Health: โIs Azure experiencing a service issue that affects me?โ
- Monitor: โWhat telemetry is my environment producing?โ
- Log Analytics: โWhat do the collected logs tell me?โ
- Alerts: โNotify or trigger an action when a condition is met.โ
- Application Insights: โHow is my application behaving?โ
Monitoring traps
- Azure Service Health is for Azure service issues and maintenance, not application debugging.
- Application Insights is for application telemetry, not policy compliance.
- Azure Monitor collects and analyzes telemetry; Azure Policy governs configurations.
- Advisor provides recommendations, not real-time incident notifications.
- Log Analytics is for querying logs; it is part of the monitoring workflow.
5. Service Selection Guide
5.1 Compute quick selection
| If the scenario says... | Think... | Avoid choosing... |
|---|---|---|
| โNeed OS-level controlโ | Azure Virtual Machines | App Service or Functions |
| โGroup of identical VMs that scalesโ | VM Scale Sets | Availability set |
| โSpread VMs across fault and update domainsโ | Availability set | VM Scale Sets |
| โRemote Windows desktops and appsโ | Azure Virtual Desktop | Ordinary VM scaling |
| โPortable app package and dependenciesโ | Containers | Resource groups |
| โTriggered code; no server managementโ | Azure Functions | Full VMs |
| โManaged hosting for a web appโ | Azure App Service | VMs unless OS control is required |
5.2 Networking quick selection
| If the scenario says... | Think... |
|---|---|
| โPrivate network address space in Azureโ | VNet |
| โDivide the VNetโ | Subnet |
| โConnect VNets privatelyโ | VNet peering |
| โResolve namesโ | Azure DNS |
| โEncrypted tunnel over the internetโ | VPN Gateway |
| โPrivate circuit; avoid public internetโ | ExpressRoute |
| โUse a private IP to reach a supported Azure serviceโ | Private endpoint |
| โReach service through public addressโ | Public endpoint |
5.3 Storage quick selection
| If the scenario says... | Think... |
|---|---|
| โImages, video, backups, unstructured objectsโ | Blob Storage |
| โSMB file shareโ | Azure Files |
| โAsynchronous messagesโ | Queue Storage |
| โNoSQL key-attribute entitiesโ | Table Storage |
| โVM operating-system or data diskโ | Managed disks |
| โFrequent accessโ | Hot tier |
| โInfrequent but online accessโ | Cool tier |
| โRare access; retrieval delay acceptableโ | Archive tier |
| โCopies within one datacenterโ | LRS |
| โCopies across zones in the primary regionโ | ZRS |
| โReplication to a secondary regionโ | GRS |
| โZones plus secondary regionโ | GZRS |
5.4 Identity and security quick selection
| If the scenario says... | Think... |
|---|---|
| โCloud identity and access managementโ | Microsoft Entra ID |
| โManaged LDAP, Kerberos, or domain joinโ | Entra Domain Services |
| โPartners use their own identityโ | External identities |
| โSign in onceโ | SSO |
| โRequire an extra verification factorโ | MFA |
| โBiometrics or security keys without passwordsโ | Passwordless |
| โRequire MFA for location, device, user, or app conditionsโ | Conditional Access |
| โAllow team to restart VMs onlyโ | Azure RBAC |
| โVerify explicitly; assume breachโ | Zero Trust |
| โLayered controlsโ | Defense in depth |
| โSecurity posture recommendations and workload protectionโ | Defender for Cloud |
5.5 Governance and operations quick selection
| If the scenario says... | Think... |
|---|---|
| โApproved region or mandatory tagโ | Azure Policy |
| โPrevent deletionโ | CanNotDelete lock |
| โPrevent changes and deletionโ | ReadOnly lock |
| โClassify by owner, environment, departmentโ | Tags |
| โGovern data estateโ | Microsoft Purview |
| โEstimate cost before deployingโ | Pricing calculator |
| โBudgets and spending trendsโ | Cost Management |
| โOptimization recommendationsโ | Azure Advisor |
| โAzure incident or maintenanceโ | Azure Service Health |
| โMetrics, logs, and alertsโ | Azure Monitor |
| โQuery logsโ | Log Analytics |
| โApplication request rates and failuresโ | Application Insights |
6. Architecture Patterns
Pattern 1: Lift-and-shift legacy workload
Scenario: A company wants to migrate a legacy server application with minimal code change and retain guest OS control.
Recommended solution: Azure Virtual Machines, usually aligned with IaaS.
Why alternatives fail:
- App Service is a managed web-hosting platform and may require application compatibility changes.
- Functions is for event-driven serverless code.
- SaaS replaces the application with a complete provider-managed application.
Pattern 2: Managed web application
Scenario: Developers want to deploy a web app without patching operating systems.
Recommended solution: Azure App Service Web Apps.
Why alternatives fail:
- A VM can host the app, but it adds OS management.
- Functions may be appropriate for event-driven components but not automatically for a conventional web application.
Pattern 3: Event-driven processing
Scenario: Run code when a message arrives, without managing servers.
Recommended solution: Azure Functions.
Why alternatives fail:
- VMs add unnecessary infrastructure administration.
- Queue Storage may store the message, but it does not execute the code.
Pattern 4: Same-region datacenter-failure protection
Scenario: Keep an application resilient to a datacenter-level failure while staying in one region.
Recommended solution: Availability zones.
Why alternatives fail:
- Availability sets address fault and update domains but are not the same as separate datacenter locations.
- Region pairs involve a separate region.
Pattern 5: On-premises-to-Azure network connection
Scenario A: Secure encrypted connection over the public internet.
Recommended solution: VPN Gateway.
Scenario B: Private connection that avoids the public internet.
Recommended solution: ExpressRoute.
Why alternatives fail:
- VNet peering connects VNets, not an on-premises site by itself.
- Azure DNS resolves names and does not create the connection.
Pattern 6: Private access to Azure Storage
Scenario: The storage account should be reachable using a private IP in a VNet.
Recommended solution: Private endpoint.
Why alternatives fail:
- A public endpoint does not meet the private-IP requirement.
- A subnet is a network segment but does not by itself create private access to the storage service.
Pattern 7: Storage choice by data shape
Scenario: Choose storage for a specific data type.
| Data shape | Choose |
|---|---|
| Objects such as images or backups | Blob |
| Shared folders over SMB | Files |
| Messages between components | Queue |
| Non-relational key-attribute entities | Table |
| VM disks | Managed disks |
Pattern 8: Governance at scale
Scenario: Enforce approved regions across multiple subscriptions.
Recommended solution: Assign Azure Policy at a management-group scope.
Why alternatives fail:
- A resource lock protects a resource from deletion or changes but does not enforce deployment rules.
- Tags classify resources but do not enforce all configuration standards.
- RBAC controls user actions, not resource configuration compliance.
Pattern 9: Monitoring a slowdown
Scenario A: Determine whether Azure has an incident affecting your resource.
Recommended solution: Azure Service Health.
Scenario B: Investigate application request rates, failures, or performance.
Recommended solution: Application Insights.
Scenario C: Query logs collected from the environment.
Recommended solution: Log Analytics.
Pattern 10: Repeatable deployments
Scenario: Create consistent environments using version-controlled files.
Recommended solution: Infrastructure as code. For a declarative JSON Azure deployment, use an ARM template.
Why alternatives fail:
- Manual portal deployment is less repeatable.
- Cloud Shell is an interface for commands, not a deployment definition.
7. Exam Traps
7.1 Misleading wording
Watch for these decisive words:
| Word or phrase | Meaning |
|---|---|
| โOver the public internetโ | VPN Gateway |
| โAvoid the public internetโ | ExpressRoute |
| โPrivate IP inside a VNetโ | Private endpoint |
| โDatacenter-level failure within a regionโ | Availability zone |
| โFault domains and update domainsโ | Availability set |
| โScale a group of identical VMsโ | VM Scale Sets |
| โWithout managing serversโ | Functions or serverless |
| โWithout patching the OSโ | PaaS-style service such as App Service |
| โGuest OS controlโ | VM / IaaS |
| โPrevent deletionโ | CanNotDelete lock |
| โPrevent modification and deletionโ | ReadOnly lock |
| โStandards and complianceโ | Azure Policy |
| โWhat can the user do?โ | RBAC |
| โService incident or maintenanceโ | Service Health |
| โRecommendationsโ | Advisor |
| โApplication requests and failuresโ | Application Insights |
7.2 Wrong-but-plausible answers
Many distractors are valid Azure services but belong to the wrong category.
Examples:
- Azure DNS is valid, but it does not connect networks.
- Azure Data Box is valid, but it does not monitor resources.
- Azure Policy is valid, but it does not copy files.
- Azure Files is valid, but it does not store asynchronous messages.
- Azure Monitor is valid, but it does not enforce allowed regions.
- Resource locks are valid, but they do not assign user permissions.
- Tags are valid, but they do not automatically lower costs.
- VNet peering is valid, but it is not an identity or authentication service.
7.3 Common distractor categories
When eliminating answers, ask whether the option belongs to the correct category:
| Category | Examples |
|---|---|
| Compute | VMs, Scale Sets, Functions, App Service, containers, Virtual Desktop |
| Networking | VNet, subnet, peering, DNS, VPN Gateway, ExpressRoute, endpoints |
| Storage | Blob, Files, Queue, Table, disks, tiers, redundancy |
| Identity | Entra ID, Domain Services, SSO, MFA, Conditional Access, RBAC |
| Governance | Policy, locks, tags, Purview, management groups |
| Deployment | Portal, Cloud Shell, CLI, PowerShell, ARM, templates, Arc |
| Monitoring | Advisor, Service Health, Monitor, Log Analytics, alerts, Application Insights |
| Cost | Pricing calculator, Cost Management, budgets |
If three answers are from unrelated categories and one matches the scenario category, the correct choice is usually clear.
7.4 Elimination strategy
Use this five-step method:
-
Name the category.
Example: โThis is a networking question.โ -
Extract the hard constraint.
Example: โThe connection must avoid the public internet.โ -
Remove wrong-category options.
Azure Policy, Blob Storage, and Application Insights cannot create the connection. -
Compare the remaining near-matches.
VPN Gateway is encrypted but uses the public internet. ExpressRoute meets the private-connectivity requirement. -
Choose the most direct answer.
Do not add services that the question does not require.
7.5 Common candidate mistakes
- Reading only the service names and ignoring decisive wording.
- Choosing the most powerful option instead of the simplest correct option.
- Confusing management hierarchy with physical architecture.
- Treating Policy, locks, RBAC, and tags as interchangeable.
- Treating Advisor, Service Health, and Monitor as interchangeable.
- Selecting ExpressRoute whenever security is mentioned, even when VPN Gateway is sufficient.
- Forgetting that IaaS VM guest OS patching is the customer's responsibility.
- Confusing a storage access tier with a redundancy option.
- Confusing Queue Storage with file storage.
- Confusing SSO with MFA.
- Forgetting that Conditional Access applies rules based on signals.
8. Quick Memory Rules
8.1 Rules of thumb
- Own the OS? Choose VM / IaaS.
- Own the code, not the OS? Choose PaaS.
- Use the complete application? Choose SaaS.
- Triggered code, no server management? Choose Functions.
- Private network in Azure? Choose VNet.
- Split a VNet? Choose subnets.
- Connect VNets? Choose peering.
- Secure tunnel over internet? Choose VPN Gateway.
- Private circuit avoiding internet? Choose ExpressRoute.
- Private IP for a service? Choose private endpoint.
- Objects? Blob. Shares? Files. Messages? Queue. Entities? Table. VM blocks? Managed disks.
- Frequent access? Hot. Occasional access? Cool. Rare access? Archive.
- One datacenter? LRS. Zones? ZRS. Geography? GRS. Zones plus geography? GZRS.
- Cloud identity? Entra ID. Managed legacy domain features? Entra Domain Services.
- One sign-in? SSO. Extra factor? MFA. No password? Passwordless. Conditional rule? Conditional Access.
- Permissions? RBAC. Standards? Policy. Protection from changes? Lock. Classification? Tags.
- Estimate future price? Pricing calculator. Analyze spending? Cost Management.
- Recommendations? Advisor. Platform incident? Service Health. Telemetry? Monitor. Logs? Log Analytics. Application behavior? Application Insights.
- GUI? Portal. Browser shell? Cloud Shell.
azcommands? CLI. Cmdlets? PowerShell. Declarative JSON? ARM template. Outside Azure? Arc.
8.2 Architecture hierarchy memory aid
Use: M โ S โ RG โ R
- Management group
- Subscription
- Resource Group
- Resource
Think: Manage subscriptions, group resources, deploy resources.
8.3 Storage redundancy memory aid
Use: L โ Z โ G โ GZ
- LRS: local datacenter.
- ZRS: zones in one region.
- GRS: geographic secondary region.
- GZRS: zones plus geographic secondary region.
8.4 Monitoring memory aid
Use: Recommend โ Health โ Monitor โ Logs โ Alert โ App
- Advisor recommends.
- Service Health reports Azure service issues.
- Monitor gathers telemetry.
- Log Analytics queries logs.
- Alerts notify.
- Application Insights explains application behavior.
9. Final Revision Notes
9.1 Highest-yield review points
Before the exam, make sure you can answer these without hesitation:
- Public versus private versus hybrid cloud.
- CapEx versus OpEx and consumption-based pricing.
- Shared responsibility in IaaS, PaaS, and SaaS.
- High availability, scalability, elasticity, reliability, predictability, governance, and manageability.
- IaaS versus PaaS versus SaaS versus serverless.
- Region versus datacenter versus zone versus region pair.
- Management group versus subscription versus resource group versus resource.
- VM versus Scale Set versus availability set versus Virtual Desktop.
- Functions versus containers versus App Service versus VMs.
- VNet, subnet, peering, DNS, VPN Gateway, ExpressRoute, public endpoints, and private endpoints.
- Blob, Files, Queue, Table, managed disks, tiers, and redundancy.
- AzCopy, Storage Explorer, File Sync, Azure Migrate, and Data Box.
- Entra ID, Entra Domain Services, external identities, SSO, MFA, passwordless, Conditional Access, and RBAC.
- Zero Trust, defense in depth, and Defender for Cloud.
- Pricing calculator, Cost Management, tags, and Advisor.
- Policy, locks, Purview, and management groups.
- Portal, Cloud Shell, CLI, PowerShell, Arc, IaC, ARM, and templates.
- Advisor, Service Health, Monitor, Log Analytics, alerts, and Application Insights.
9.2 Last-day revision list
Do these in order:
- Recite the architecture hierarchy.
- Recite storage service mappings.
- Recite storage redundancy mappings.
- Compare VPN Gateway and ExpressRoute.
- Compare Policy, locks, RBAC, and tags.
- Compare Advisor, Service Health, Monitor, Log Analytics, alerts, and Application Insights.
- Compare Entra ID and Entra Domain Services.
- Compare SSO, MFA, passwordless, and Conditional Access.
- Compare VMs, App Service, containers, and Functions.
- Review every question you previously answered incorrectly.
9.3 One-minute rapid review
- Guest OS control โ VM.
- Managed web app โ App Service.
- Triggered code โ Functions.
- Two VNets โ peering.
- Internet tunnel โ VPN Gateway.
- Private circuit โ ExpressRoute.
- Private service IP โ private endpoint.
- Objects โ Blob.
- SMB โ Files.
- Messages โ Queue.
- NoSQL entities โ Table.
- VM disks โ managed disks.
- Standards โ Policy.
- Prevent deletion โ CanNotDelete lock.
- Prevent all changes โ ReadOnly lock.
- Permissions โ RBAC.
- Estimate cost โ pricing calculator.
- Spending trends and budgets โ Cost Management.
- Recommendations โ Advisor.
- Azure incident โ Service Health.
- Telemetry โ Monitor.
- Query logs โ Log Analytics.
- Application telemetry โ Application Insights.
10. Exam-Day Checklist
Must-know topics
Cloud concepts
- I can distinguish public, private, and hybrid cloud.
- I can explain CapEx, OpEx, and consumption-based pricing.
- I understand shared responsibility across IaaS, PaaS, and SaaS.
- I can distinguish scalability, elasticity, high availability, and reliability.
- I can select IaaS, PaaS, SaaS, or serverless for a scenario.
Architecture and services
- I know region, datacenter, availability zone, region pair, and sovereign region.
- I know the hierarchy: management group โ subscription โ resource group โ resource.
- I can select VM, Scale Set, availability set, Virtual Desktop, container, Function, or App Service.
- I know that a NIC connects a VM to a VNet and managed disks store VM OS and data disks.
- I can select VNet, subnet, peering, DNS, VPN Gateway, ExpressRoute, public endpoint, or private endpoint.
- I can select Blob, Files, Queue, Table, or managed disks.
- I can select Hot, Cool, or Archive tiers.
- I can distinguish LRS, ZRS, GRS, and GZRS.
- I know AzCopy, Storage Explorer, File Sync, Azure Migrate, and Data Box.
- I can distinguish Entra ID and Entra Domain Services.
- I can distinguish SSO, MFA, passwordless, Conditional Access, and RBAC.
- I understand Zero Trust, defense in depth, and Defender for Cloud.
Management and governance
- I can select the pricing calculator, Cost Management, tags, or Advisor.
- I can distinguish Policy, locks, RBAC, and tags.
- I know the difference between CanNotDelete and ReadOnly locks.
- I know Microsoft Purview is a data-governance capability.
- I can select portal, Cloud Shell, CLI, PowerShell, Arc, IaC, ARM, or ARM templates.
- I can distinguish Advisor, Service Health, Monitor, Log Analytics, alerts, and Application Insights.
Final confidence checklist
- I read the entire question and identify the primary requirement before looking at the options.
- I eliminate answers from the wrong service category.
- I choose the simplest service that directly meets the requirement.
- I explain why the strongest distractor is wrong.
- I do not confuse a management tool with a network, storage, compute, or identity service.
- I do not confuse a storage tier with a redundancy option.
- I do not confuse an Azure platform incident with an application-performance issue.
- I do not assume a more expensive or complex option is automatically better.
- I am ready to use the practice bank for mixed review.
Appendix A โ Confusing-Service Master Table
| Confusing services | Correct distinction |
|---|---|
| IaaS vs PaaS vs SaaS | OS control vs managed app platform vs complete application |
| VM vs Functions | OS-level compute vs serverless event-driven code |
| VM Scale Sets vs availability sets | Scaling similar VMs vs distributing VMs across fault and update domains |
| Availability set vs availability zone | Host-domain distribution vs physically separate location within a region |
| Availability zone vs region pair | Same-region datacenter isolation vs cross-region relationship |
| VNet peering vs VPN Gateway vs ExpressRoute | VNet-to-VNet private connection vs encrypted internet tunnel vs private circuit |
| Public endpoint vs private endpoint | Public address vs private IP in a VNet |
| Blob vs Files | Object storage vs managed file shares |
| Queue vs Table | Messages vs non-relational entities |
| Hot vs Cool vs Archive | Frequent vs infrequent online vs rare delayed access |
| LRS vs ZRS vs GRS vs GZRS | Local vs zones vs geography vs zones plus geography |
| Entra ID vs Entra Domain Services | Cloud identity vs managed legacy domain capabilities |
| SSO vs MFA vs passwordless vs Conditional Access | One login vs extra factor vs no password vs rule-based access enforcement |
| RBAC vs Policy vs locks vs tags | Permissions vs standards vs protection from changes vs metadata |
| Pricing calculator vs Cost Management | Estimate proposed design vs analyze spending and budgets |
| Advisor vs Service Health vs Monitor | Recommendations vs Azure platform issues vs telemetry platform |
| Monitor vs Log Analytics vs Application Insights | Telemetry platform vs log queries vs app performance monitoring |
| Portal vs Cloud Shell vs CLI vs PowerShell | Browser GUI vs browser shell vs az commands vs cmdlets |
| Azure Migrate vs Azure Data Box | Migration assessment and planning vs offline bulk data transfer |
| ARM vs ARM template | Azure management layer vs declarative JSON deployment file |
Appendix B โ Source Alignment Note
This course was synthesized from the analyzed AZ-900 practice-question bank and aligned with the Microsoft Learn AZ-900 study guide for skills measured as of January 14, 2026. The guide intentionally consolidates repeated questions into original explanations, decision frameworks, and service-selection rules. It is a revision course, not a reproduction of live exam content.
Unlock the full course
All 51 modules with detailed explanations, code examples, and exam tips.
A project team is documenting a foundational decision. Coho Winery wants to analyze Azure spending, create budgets, and review cost trends. Which capability is the most appropriate? The choice must address the main requirement rather than a secondary benefit.
This Question is Locked
You're viewing 35 of 1100 free questions.
trending_up Certified pros earn 20-30% more
Higher salary: IT certifications add $12,000-$25,000/year on average to your paycheck
Job security: 87% of hiring managers prefer candidates with certifications : you become irreplaceable
More opportunities: Freelance gigs, remote roles, and promotions open up instantly
Practice all questions: Comprehensive practice is the #1 predictor of passing
Real Exam : Upgrade to Unlock
Available in Q&A + Course + Mock Exam package
You've already started : one exam away from a career upgrade.