AZ-104 Exam Preparation Guide
Transcript analysis and practical study plan for preparing for the Microsoft AZ-104: Microsoft Azure Administrator exam.
Supporting AZ-104 video
Squarespace is blocking the embedded player on this plan, so this uses a clickable thumbnail instead. Select the image below to open the video in YouTube.

1. Executive summary
- Learn the official skills list first. Avoid spending most of your time on adjacent topics that are not in the AZ-104 outline.
- Practise in a real Azure subscription or Microsoft Learn sandbox where possible. The exam expects operational judgement.
- Treat each question as a scenario: identify the required service, scope, permissions, constraints and best-practice answer.
- Use least privilege, cost control, security, resilience and operational maintainability as decision filters.
- Use the Exam Sandbox before exam day so the interface and question styles do not cost you time.
2. Exam blueprint and weighting
The five objective domains below should be used as your revision structure.
| Item | What to know |
|---|---|
| Exam | AZ-104: Microsoft Azure Administrator |
| Certification | Microsoft Certified: Azure Administrator Associate |
| Audience | Azure administrators implementing, managing and monitoring Azure environments. |
| Prerequisite working knowledge | Operating systems, networking, servers, virtualisation, PowerShell, Azure CLI, Azure portal, ARM/Bicep and Microsoft Entra ID. |
| Pass score | 700 or greater. |
| Main preparation tools | Official study guide, Microsoft Learn modules, hands-on labs, free Practice Assessment and Exam Sandbox. |
| Domain | Weight | Priority | Why it matters |
|---|---|---|---|
| Manage Azure identities and governance | 20-25% | Very high | Covers Entra users/groups, RBAC, subscriptions, policy, locks, tags, cost management and management groups. These are foundational across almost every Azure environment. |
| Implement and manage storage | 15-20% | High | Tests access, storage accounts, redundancy, SAS, Azure Files, Blob Storage, lifecycle management, soft delete and versioning. |
| Deploy and manage Azure compute resources | 20-25% | Very high | Covers ARM/Bicep, VMs, VM availability, scale sets, containers, App Service and scaling. This is one of the most task-heavy areas. |
| Implement and manage virtual networking | 15-20% | High | Covers VNets, subnets, peering, UDRs, NSGs, ASGs, Bastion, endpoints, DNS and load balancers. Many questions are troubleshooting or rule-evaluation scenarios. |
| Monitor and maintain Azure resources | 10-15% | Medium, but do not skip | Covers Azure Monitor, logs, alerts, insights, Network Watcher, backups and Site Recovery. Smaller weighting, but procedural details can be decisive. |
Supporting links
3. Preparation strategy
3.1 The working method
- Read the current official Microsoft Learn AZ-104 study guide and copy the skills list into a tracker.
- For each skill, decide whether you can perform it in the Azure portal, explain when to use it, and identify the common exam trap.
- Build small labs for the key tasks. Do not only watch videos.
- Use the free Practice Assessment to find weak areas. Treat it as diagnostic rather than as the exam itself.
- Use the Exam Sandbox before booking or at least before exam week.
- Do a final pass on high-yield comparison topics: RBAC vs Entra roles, policy vs locks, service endpoints vs private endpoints, scale up vs scale out, ACI vs ACA vs AKS, Backup vs Site Recovery.
3.2 What the exam is really testing
- Scope: management group, subscription, resource group or resource.
- Permissions: which built-in role is enough, and where it should be assigned.
- Constraints: region support, resource move limits, networking rules, SKU/tier limitations and feature prerequisites.
- Operational sequence: which steps must happen first, especially for backup, recovery, networking and deployment scenarios.
- Best practice: least privilege, policy-based governance, cost controls, secure access and resilient architecture.
3.3 Topics not to over-prioritise
The transcript warns candidates not to chase areas that are not in the official skills list. You may need awareness of related features, but deep configuration of the following should not dominate AZ-104 revision:
- Deep Intune administration.
- Deep Conditional Access design, beyond knowing licensing relationships where relevant.
- Advanced Microsoft Entra Privileged Identity Management configuration, beyond recognising P2 licensing where relevant.
- Deep Kubernetes design. Know when AKS is appropriate, but the AZ-104 container focus is ACR, ACI and Azure Container Apps.
- Application development internals. For App Service and Application Insights, focus on administrator tasks.
Supporting links
4. Domain 1 - Manage Azure identities and governance
Weighting: 20-25%This is one of the highest-value revision areas.
4.1 Core knowledge
| Area | You need to know |
|---|---|
| Microsoft Entra users and groups | Create and manage users and groups; understand cloud identities, directory-synchronised identities and guest users; manage user and group properties; understand group-based access. |
| Licensing | Understand Free, P1, P2 and Entra Suite at a high level; know that some capabilities require premium licensing; know how licence assignment and group inheritance work. |
| External users | Guest users are added to your tenant as external identities. Know invitation flow, access review concepts and sponsor/review best practice. |
| SSPR | Know that self-service password reset is in scope and understand its purpose, enablement and user impact. |
| Azure RBAC | Know Owner, Contributor and Reader well. Understand role definitions, role assignments, security principals and scopes. |
| Entra roles vs Azure RBAC roles | Entra roles control directory administration. Azure RBAC controls Azure resource access. Do not confuse the two. |
| Governance | Azure Policy, initiatives, resource locks, tags, resource groups, subscriptions, costs, budgets, Advisor recommendations and management groups. |
4.2 Exam traps and decision rules
- Permissions are additive across assignments unless denied by other controls. If a user is Reader at one scope and Contributor at a lower scope, evaluate the effective result at the target resource.
- Scope inheritance flows downwards: management group to subscription to resource group to resource.
- Least privilege is normally the expected answer. Assign the narrowest role at the narrowest useful scope.
- Policy and RBAC do different things. RBAC says who can attempt an action. Policy can restrict what is allowed even when a user has permission.
- Resource locks protect against accidental deletion or modification. Policy enforces standards at scale. Know when to use each.
- Every Azure resource belongs to exactly one resource group; resource groups cannot be nested or renamed; resources in a group can be in different regions.
- Management groups sit above subscriptions and are used to organise policy and access control across multiple subscriptions.
- Cost questions often involve budgets, alerts, Azure Advisor recommendations, reservations and Azure Hybrid Benefit.
Supporting links
5. Domain 2 - Implement and manage storage
Weighting: 15-20%The topic list is broad, so do not underestimate this section.
5.1 Core knowledge
| Area | You need to know |
|---|---|
| Storage access | Storage firewalls, virtual network rules, public IP restrictions, trusted Azure services, service endpoints and private endpoints. |
| SAS tokens | User delegation SAS, service SAS and account SAS. Prefer user delegation SAS where possible because it uses Microsoft Entra credentials and RBAC. |
| Stored access policies | Extra control for SAS tokens; understand revocation by deleting or changing the policy and that propagation can take a short time. |
| Access keys | Know why hard-coding keys is poor practice; use Key Vault and understand the effect of disabling shared key access. |
| Azure Files | SMB/NFS access patterns, identity-based access for SMB scenarios, Windows ACL preservation, and the need for TCP 445 for SMB connectivity. |
| Storage redundancy | LRS, ZRS, GRS, GZRS and read-access variants. Know the trade-off between cost, availability, regional disaster protection and read access. |
| Encryption | Storage Service Encryption is automatic. Know Microsoft-managed keys versus customer-managed keys and Key Vault dependency for customer-managed keys. |
| Tools | Storage Explorer is GUI-based. AzCopy is command-line and optimised for copying data to/from Azure Storage. Know authentication options and use cases. |
| Blob and file protection | Soft delete, snapshots, lifecycle management, storage tiers and blob versioning. |
5.2 High-yield details from the transcript
- Storage firewalls and virtual network rules apply to data-plane operations.
- A stored access policy can make SAS revocation easier than managing individual SAS tokens directly.
- For Azure file shares over SMB, port 445 must be open and SMB 3.0 encryption matters for secure transfer.
- Blob access levels are private, blob and container. Know the security implication of each.
- Blob tiers: hot, cool, cold and archive. The transcript highlights minimum retention periods of 30 days for cool, 90 days for cold and 180 days for archive.
- Archive has the lowest storage cost but rehydration takes time. Moving data before minimum retention does not block the move, but may create early deletion charges.
- Blob lifecycle policies automate tiering and deletion based on rules such as blob age, container or path.
- Blob versioning creates versions on modification; understand the cost and latency implications of many versions.
Supporting links
6. Domain 3 - Deploy and manage Azure compute resources
Weighting: 20-25%Along with identities and governance, this is a top-priority domain.
6.1 Core knowledge
| Area | You need to know |
|---|---|
| ARM templates | JSON structure: schema, contentVersion, parameters, variables, functions, resources and outputs. Be able to interpret, modify, deploy and export templates. |
| Bicep | Readable domain-specific language for Azure deployment. Know parameters, conditions, loops, modules, deployment and conversion/compilation to ARM. |
| Virtual machines | Create VMs, select image/size, configure network, disks, management options, extensions, backups, monitoring and auto-shutdown. |
| VM sizing and disks | Know VM family intent, disk attachment, data disk limits by size, managed disks and encryption at host. |
| Availability | Availability zones, availability sets, update domains, fault domains, load balancers and VM scale sets. |
| Moving VMs | Moving between resource groups/subscriptions versus region moves; understand dependencies and downtime impact. |
| Container Registry | Private Docker-compatible registry for container images; know tiers, image storage and integration with AKS, App Service and other services. |
| ACI | Fastest simple container runtime without managing VMs; good for simple or burst workloads. |
| Azure Container Apps | Managed serverless container platform for microservices and event-driven/containerised workloads. |
| App Service | App Service plan, app creation, TLS/certificates, custom DNS, backup, networking, scaling and deployment slots. |
6.2 Exam traps and decision rules
- Do not memorise every VM size. Know the family pattern and match the workload: general purpose, compute optimised, memory optimised and so on.
- Scale up means increasing the power/tier/size of an instance. Scale out means adding instances.
- VM scale sets manage groups of identical VMs and support automatic scale behaviour.
- ACI is for quick simple container execution; Azure Container Apps is managed/serverless for containerised apps; AKS is the full orchestration answer when the scenario requires Kubernetes-level control.
- An App Service plan defines the compute resources and cost model. Multiple apps can share the same plan.
- Deployment slot settings can be marked as slot settings so they stick to the slot and do not swap.
- For ARM/Bicep questions, focus on reading structure and intent rather than memorising full syntax.
Supporting links
7. Domain 4 - Implement and manage virtual networking
Weighting: 15-20%Expect scenario and troubleshooting questions.
7.1 Core knowledge
| Area | You need to know |
|---|---|
| VNets and subnets | Create VNets/subnets; understand address spaces, segmentation, communication between Azure resources, internet access and hybrid connectivity. |
| VNet peering | Connect VNets; know peering status, virtual network access, forwarded traffic and gateway transit options. |
| Public IP addresses | Configure and understand static/dynamic allocation and SKU implications for modern deployments. |
| User-defined routes | Route tables, next hop types, overriding system routes and using network virtual appliances. |
| Network troubleshooting | VM-to-VM, VM-to-internet, secondary NIC and route problems. Know what to check first. |
| NSGs | Inbound/outbound rules, default rules, priority order and five-tuple evaluation: source, source port, destination, destination port and protocol. |
| ASGs | Logical grouping of VMs for NSG rule targeting. ASGs do not filter traffic by themselves; NSGs do. |
| Azure Bastion | Secure RDP/SSH through browser over TLS without exposing VM public IP addresses. |
| Service endpoints and private endpoints | Both secure access to PaaS, but service endpoints extend subnet identity to the service while private endpoints place a private IP for the service in the VNet. |
| DNS and load balancing | Azure DNS vs Private DNS; internal vs public Load Balancer; frontend IPs, backend pools, rules and health probes. |
7.2 Exam traps and decision rules
- For NSG questions, sort rules by priority and direction before deciding allow/deny.
- Remember default NSG rules. A custom higher-priority rule can override the default behaviour.
- ASGs simplify rule management when IP addresses change; they do not replace NSGs.
- UDRs are used when the scenario wants to force traffic through a firewall or network virtual appliance.
- Private endpoints are usually the answer when the requirement is private IP access to a PaaS service.
- For Bastion scenarios, look for requirements such as no VM public IP, browser-based RDP/SSH and reduced attack surface.
- For load balancer troubleshooting, check frontend IP, backend pool membership, rules, health probes, NSGs and VM health.
Supporting links
8. Domain 5 - Monitor and maintain Azure resources
Weighting: 10-15%This is the smallest domain, but backup and recovery questions often require exact steps.
8.1 Core knowledge
| Area | You need to know |
|---|---|
| Azure Monitor metrics | Interpret metrics such as CPU and IOPS; understand near-real-time monitoring and data sources. |
| Logs and Log Analytics | Configure diagnostic/log settings, query and analyse logs, save queries and create dashboards. |
| Alerts | Alert rules include scope, condition/criteria, severity, action groups and alert processing rules. |
| Monitor Insights | Use insights for VMs, storage accounts and networks; understand what extra visibility each provides. |
| Application Insights | Know the administrator-level purpose: request rates, response times, failures, dependencies, exceptions, page views and custom events. |
| Network Watcher | Topology, diagnostics, Connection Monitor, performance/connectivity checks and network troubleshooting. |
| Recovery Services vault | Used for backup and recovery scenarios; stores recovery points and supports workloads such as Azure VMs and Azure Files. |
| Backup vault and policies | Know backup schedules, retention, redundancy choices and reporting. |
| Azure Site Recovery | Disaster recovery and replication/failover to a secondary Azure region. Know test failover and failover sequence at a high level. |
| Reporting and alerts | Backup integrates with Resource Graph, Monitor alerts/logs and built-in/custom reporting. |
8.2 Exam traps and decision rules
- Metrics and logs are different data types. Metrics are numerical time-series values; logs are queryable event/detail records.
- Log Analytics is used when you need richer retention, KQL querying, saved queries and dashboards.
- Action groups define what happens when an alert fires.
- Network Watcher is the expected diagnostic service for Azure networking problems.
- Azure Backup is for backup/restore. Azure Site Recovery is for disaster recovery and failover.
- For restore or failover questions, look carefully for required order: permissions, vault, policy/replication, validation, test failover/restore, cleanup.
Supporting links
9. Practical lab checklist
Use this as a practical build plan. The aim is not to create a large production environment, but to touch the administrator tasks likely to appear in scenarios.
| Lab | Tasks to perform | Evidence you should be able to explain |
|---|---|---|
| 1. Governance baseline | Create resource groups; add tags; create a budget; assign a simple policy or initiative; apply a lock. | Policy vs lock; tag inheritance limitations; cost alert permissions; resource group constraints. |
| 2. Entra and RBAC | Create a test user/group; assign Reader at subscription and Contributor at a resource group; review effective access. | RBAC inheritance, least privilege, Azure RBAC vs Entra admin roles. |
| 3. Storage security | Create a storage account; restrict access with firewall/VNet rule; create SAS; rotate keys; test Storage Explorer/AzCopy. | SAS types, stored access policy use, firewall scope, key security. |
| 4. Blob lifecycle | Upload blobs; configure hot/cool/archive tiering; enable soft delete and versioning; create lifecycle rule. | Tier trade-offs, retention periods, lifecycle rule conditions, version cost. |
| 5. Azure Files | Create a file share; connect from a Windows client; test SMB connectivity. | Port 445, SMB versions, share access, snapshot/soft delete. |
| 6. ARM/Bicep deployment | Deploy a simple storage account or VNet using Bicep; inspect equivalent ARM JSON. | Parameters, resources, outputs, deployment flow. |
| 7. VM operations | Create a VM; add a data disk; resize VM; configure monitoring; enable backup. | VM size implications, disks, backup flow, management options. |
| 8. Availability and scale | Create an availability set or scale set; review fault/update domains or scaling options. | Availability choices, scale up vs scale out, VMSS behaviour. |
| 9. Containers and App Service | Create ACR; run a container in ACI or Azure Container Apps; create App Service plan/app; configure deployment slot. | ACI vs ACA vs AKS; App Service plan cost/scaling; sticky slot settings. |
| 10. Networking | Create two VNets; peer them; add NSGs and ASGs; configure UDR; deploy Bastion if possible. | Peering settings, NSG priority, ASG purpose, UDR next hop, Bastion architecture. |
| 11. Load balancing | Create a public or internal Load Balancer with backend VMs; test health probes. | Frontend/backend/rule/probe relationship and troubleshooting order. |
| 12. Monitoring and recovery | Create Log Analytics workspace; configure diagnostic settings; create alert/action group; test backup and Site Recovery concepts. | Metrics vs logs, alerts, Network Watcher, Backup vs Site Recovery. |
Supporting links
10. Exam-day strategy
- Read the requirement before the scenario text. Identify the task: deploy, secure, troubleshoot, reduce cost, monitor or recover.
- Look for scope words: tenant, management group, subscription, resource group, resource, subnet, NIC, storage account, container or file share.
- Eliminate answers that are too broad, too privileged, too expensive or do not meet the stated constraint.
- When two answers work, prefer the one that is least privileged, policy-driven, secure and operationally simpler.
- For network questions, map direction and path: source, destination, protocol, port, route, NSG, endpoint and load balancer health.
- For sequence questions, identify dependencies: create vault before backup; define policy before assignment; create backend pool/probe/rules before expecting load balancer success.
- Use Microsoft Learn during the exam only for targeted lookup. The timer continues, so do not plan to research every question.
- Use mark-for-review strategically. Do not get stuck on one complex scenario at the expense of easier marks.
- Be careful with breaks. Microsoft warns that after starting an unscheduled break you cannot return to questions already viewed before the break.
Supporting links
11. Final readiness checklist
| Readiness item | Target before booking |
|---|---|
| Official objectives | You can explain every bullet in the current skills measured list at a high level. |
| Portal confidence | You can find where to configure each major service without searching externally. |
| Command-line confidence | You recognise common Azure CLI and PowerShell approaches for core tasks. |
| ARM/Bicep confidence | You can read a simple deployment file and identify parameters, resources and outputs. |
| Lab evidence | You have completed most of the practical labs in section 9 or equivalent Microsoft Learn labs. |
| Practice Assessment | You have used the free Practice Assessment to identify weak domains and retested after revision. |
| Exam Sandbox | You have used the sandbox and understand the interface, review screen and question interactions. |
| High-yield comparisons | You can explain the comparison topics below without notes. |
Azure RBAC vs Entra roles; Azure Policy vs resource locks; management groups vs subscriptions vs resource groups; service endpoints vs private endpoints; NSGs vs ASGs; scale up vs scale out; availability sets vs zones vs scale sets; ACI vs Azure Container Apps vs AKS; storage access keys vs SAS vs Microsoft Entra authentication; LRS/ZRS/GRS/GZRS and read-access variants; metrics vs logs; Azure Backup vs Azure Site Recovery.
Supporting links
12. Useful official links and reference library
Use these links as the reference library for the guide. The first two links are the most important to check before booking because exam objectives and certification guidance can change.