AZ-305 Solutions Architect Expert Study Guide
The Microsoft Azure AZ-305 exam tests whether you can design complete Azure infrastructure solutions for large organisations, including decisions about cloud adoption, security, governance and cost. Compared with the AZ-104 exam, AZ-305 asks you to weigh trade-offs, scalability and long-term consequences rather than carry out tasks.
This guide covers the prerequisites, the exam format, each domain, and how to prepare for the case study questions.
What AZ-305 Actually Tests
AZ-305 does not test whether you can click through the Azure Portal or run specific commands. It tests whether you can design solutions that balance performance, security, cost and compliance.
The scenarios read like real projects. A healthcare company moving to the cloud needs HIPAA compliance and disaster recovery. A financial services firm requires multi-region high availability with sub-second failover. A startup wants to minimise costs while keeping the ability to scale. You recommend the Azure services and architecture patterns that fit.
You need to know when to use Azure SQL Database or Cosmos DB, when hub-and-spoke topology matters, how to design disaster recovery to meet RTO and RPO targets, and how to apply governance across multiple teams and subscriptions.
Prerequisites: AZ-104 Is Required for the Certification
Anyone can sit the AZ-305 exam. Microsoft awards the Azure Solutions Architect Expert certification only to people who also hold Azure Administrator Associate (AZ-104), so passing the exam alone is not enough.
AZ-305 assumes you already know virtual networks, storage accounts, identity and access management, and basic resource deployment. It will not explain how to create a virtual machine or configure a network security group. It asks how you would combine those components in a design.
If you have not passed AZ-104, do that first. Experienced IT professionals typically need 2-4 weeks for it. After AZ-104, plan on 6-12 weeks of AZ-305 preparation.
Who Should Take This Exam
AZ-305 is aimed at cloud architects and senior infrastructure professionals with substantial Azure experience, including:
- Cloud Solutions Architects designing enterprise migrations
- Senior Cloud Engineers moving into architecture roles
- Infrastructure Leads planning multi-year cloud strategies
- Enterprise Architects evaluating Azure adoption
- Solutions Architects at consulting firms designing customer solutions
- Technical Decision Makers building large-scale cloud systems
You should have at least 12-18 months of hands-on Azure experience. Documentation and videos will not substitute for practical work on network design, storage decisions, security and day-to-day operations.
Candidates who do well have usually worked on several Azure projects, chosen between services themselves, and seen how those choices held up six months or a year later.
Exam Format and Structure
AZ-305 is built around case studies, which makes it different from most multiple-choice exams.
Case Study Format
Each case study is a scenario document (typically 6-8 pages) describing a company, its requirements, current infrastructure, compliance needs, budget constraints and growth plans. After reading it, you answer 8-12 multiple-choice questions about that case study.
The important details are spread through the document. A compliance requirement can sit in the middle of a paragraph, and a cost constraint may be mentioned only once. Read the whole scenario before answering, and note the details as you go.
Time Allocation
You have 100 minutes to answer the questions, within about 120 minutes of total seat time. With several case studies to get through, budget your time: around 15 minutes to read a case study and 10-15 minutes to answer its questions is reasonable. Do not let one difficult question eat into the rest.
Question Types
Questions are not all a single format. Alongside standard multiple-choice items, AZ-305 case studies can include other formats such as drag-and-drop, build-list and active-screen questions. Microsoft does not publish the exact mix of question types for a specific exam in advance. Some ask for the best solution. Others ask which requirement a proposed solution meets or fails to meet.
Five Core Exam Domains
Microsoft splits AZ-305 into five domains. Use the weightings to decide where your study time goes.
| Domain | Percentage | Key Topics |
|---|---|---|
| Design identity, governance, and monitoring solutions | 25-30% | Microsoft Entra ID, RBAC, policy, cost management, Log Analytics |
| Design data storage solutions | 25-30% | SQL Database, Cosmos DB, storage accounts, data lakes, caching |
| Design business continuity solutions | 10-15% | Backup, disaster recovery, RTO/RPO, failover strategies |
| Design infrastructure solutions | 25-30% | Virtual networks, load balancing, app deployment, hybrid connectivity |
| Design migrations | 5-10% | Assessment, migration strategies, legacy system integration |
Most of the weight sits with identity/governance, data storage and infrastructure. Business continuity still needs study, but it carries a smaller share. azureprep.com has practice tests for each domain, so you can work on weak areas before exam day.
Domain 1: Design Identity, Governance, and Monitoring Solutions (25-30%)
This domain covers Microsoft Entra ID (formerly Azure AD), role-based access control, policy enforcement, cost management and operational monitoring.
Microsoft Entra ID and Identity Architecture
Identity gets harder as an organisation grows. Designs have to support thousands of users, multiple applications, third-party integrations and security requirements that change over time.
Concepts to know:
- Directory design: single versus multiple tenants
- Hybrid identity scenarios: on-premises AD synchronised with Microsoft Entra ID
- B2B and B2C scenarios for external user collaboration
- Conditional Access policies enforcing device compliance and location restrictions
- Multi-factor authentication requirements for sensitive operations
A typical question: "A financial services company has 15,000 employees across four countries. They use on-premises Active Directory and are expanding to Azure. They require conditional access that blocks access from non-managed devices. How should they design their identity solution?"
A good answer covers directory synchronisation (Microsoft Entra Connect), conditional access policies in Microsoft Entra ID, device compliance requirements, and possibly Microsoft Entra ID P1/P2 licensing for the advanced features.
Governance and Compliance
Governance keeps Azure use consistent, secure and within budget across teams.
Azure Policy stops teams creating non-compliant resources. Management Groups apply governance hierarchically across multiple subscriptions. Template Specs and Deployment Stacks define standard environments for teams to deploy (Azure Blueprints retires on 31 January 2027, with a phased retirement beginning 31 July 2026; Microsoft recommends migrating to Deployment Stacks, with Template Specs also supported).
The hardest questions here ask you to balance control against speed. Policies that are too strict slow development down, and policies that are too loose create security and compliance risks.
Cost Management and Monitoring
Azure Monitor, Log Analytics and Azure Cost Management show how resources are used and what they cost.
Design choices drive the bill. The choice of VM family, storage tier or service SKU can multiply a monthly bill, so you need to know the cost drivers and recommend designs that meet the requirements without overspending.
Monitoring questions often present a scenario: "A development team is experiencing unexpected Azure bills. They have 200 VMs running across multiple resource groups. How would you implement monitoring to identify cost anomalies?"
The answer uses Azure Cost Management for cost analysis, Azure Monitor for resource metrics, and possibly Azure Advisor for optimisation recommendations.
Domain 2: Design Data Storage Solutions (25-30%)
Many candidates find this the hardest domain to reason about. The wrong database for a critical application causes problems that last for years.
Relational vs. NoSQL vs. Analytical Stores
The trade-offs to know:
- Azure SQL Database: ACID transactions, relational schema, strong consistency, mature ecosystem
- Cosmos DB: multi-region replication, flexible schema, eventual consistency, massive scale
- Azure Synapse Analytics: petabyte-scale data warehouse for analytical queries
- Data Lake Storage: unstructured data, big data processing, cost-effective storage
A classic question: "An e-commerce company processes millions of orders daily. They need to query order history, customer profiles, and payment information transactionally. Where should they store data?"
Answer: Azure SQL Database with suitable scaling (elastic pools or Hyperscale). Relational consistency, transactions and querying matter more here than horizontal scale.
Different scenario: "A real-time gaming company stores player stats, match results, and social connections. Data access patterns vary widely, and they need sub-millisecond latency with millions of concurrent players globally."
Answer: Cosmos DB with multi-region replication. Flexible schema, eventual consistency and global scale matter more here than strict transactions.
Caching and Performance
Azure Cache for Redis sits between applications and databases and can reduce database load by 10-100x for read-heavy workloads.
Know when caching helps and when it causes problems. It suits data that rarely changes (product catalogues, user preferences) and causes consistency issues for data that changes often (account balances, inventory counts).
Data Migration and Integration
A storage design also has to say how the data arrives. Azure Data Factory moves data from legacy systems, and Azure Synapse Pipelines integrate with data lakes. You need to know what these tools are for, not how to operate them.
Domain 3: Design Business Continuity Solutions (10-15%)
Business continuity is about keeping services running through failures. The two metrics that drive every design are RTO (Recovery Time Objective) and RPO (Recovery Point Objective).
RTO is how long the service can be down. An e-commerce site might have an RTO of 5 minutes, and an internal HR system an RTO of 8 hours.
RPO is how much data can be lost. A payment system might have an RPO of seconds, while a data warehouse might accept hours.
High Availability Architecture
High availability keeps services running when a component fails. Availability Zones (independent datacentres within a region) protect against a datacentre outage, load balancers spread traffic across instances, and managed services such as Azure App Service handle failover automatically.
A solution with RTO of minutes and RPO of minutes typically requires:
- Multiple instances across Availability Zones
- Automated health checks and failover
- Data replication with near-zero lag
Disaster Recovery Architecture
Disaster recovery covers the loss of a whole region. Azure Paired Regions provide geo-redundancy: critical data is replicated to a secondary region, and recovery runbooks are kept ready for failover.
RTO and RPO determine strategy:
- RTO/RPO of hours: backup and restore from secondary region (cheapest)
- RTO/RPO of minutes: active-active replication (most expensive)
- RTO/RPO of seconds: active-active with continuous replication (very expensive)
The exam checks that you match the approach to the stated RTO/RPO without building (and paying for) more than the scenario needs.
Domain 4: Design Infrastructure Solutions (25-30%)
This domain covers networking, compute and deployment patterns.
Virtual Network Architecture
Most organisations use hub-and-spoke topology:
- Hub: central virtual network containing shared services (firewalls, VPN gateways, DNS)
- Spokes: workload-specific virtual networks connected to hub via peering
- Central control plane: policies and monitoring through hub
The design scales to hundreds of spokes, keeps security boundaries in place, and is simpler to manage.
Landing zones formalise the pattern. Azure publishes landing zone reference architectures that teams use to create standard environments for new workloads.
Load Balancing and Traffic Management
Azure Load Balancer (Layer 4) distributes traffic across VMs. Application Gateway (Layer 7) understands HTTP and can route by hostname or path. Azure Front Door provides global load balancing with DDoS protection.
The choice depends on whether traffic is internal or external, Layer 4 or Layer 7, and regional or global.
Compute and Deployment
Virtual Machines give the most control and the most operational work. Container instances run containers without managing servers. App Service is platform-as-a-service hosting for web apps. Azure Functions runs serverless workloads.
Pick the simplest service that meets the requirements. If App Service works, use it instead of VMs. If a managed database works, do not self-host a database on VMs.
Design Patterns
Three patterns come up repeatedly on AZ-305:
Hub-and-Spoke Networking
The hub holds the shared services and security controls: firewalls, VPN gateways and DNS servers. Spokes are isolated workload networks. Traffic between spokes goes through the hub rather than directly.
Benefits: centralised security, scale to hundreds of workloads, shared services.
Drawbacks: extra latency (all traffic passes through the hub), and the hub becomes a single point of failure.
Landing Zone Design
A landing zone is a pre-configured Azure environment ready for workloads. Organisations create landing zones for different teams (development, production, data analytics) so teams can deploy on their own while governance stays in place.
Main elements: network topology, identity integration, policy enforcement, monitoring, cost allocation.
With a good landing zone in place, teams can deploy a workload in hours instead of weeks.
High Availability and Disaster Recovery
Availability Zones replicate across datacentres within a region (99.99% SLA for zone-redundant VM deployments). Paired Regions provide geo-redundancy across regions for disaster recovery; the SLA for a paired-region design depends on the specific service and replication option used, so check the target service's own SLA rather than assuming one figure.
Assume individual components will fail and design so the service keeps running when they do.
The Hardest Questions: Balancing Requirements
The hardest questions set requirements against each other and ask for the solution that meets all of them at the lowest cost.
Example: "A company needs a database solution that supports complex transactions, scales to 10 million daily requests, costs less than $2,000 monthly, and provides disaster recovery. Which solution?"
This tests whether you understand:
- SQL Database costs increase with DTU/vCore
- Cosmos DB scales horizontally but costs accumulate quickly
- Some combinations are impossible (you cannot scale SQL Database infinitely and stay below $2,000)
- The correct answer trades off capabilities
You might recommend Azure SQL Database with read replicas for some workloads, or Cosmos DB with careful partitioning for others. Either way, the answer is a trade-off you can justify.
Study Approach for a Design Exam
AZ-305 needs a different kind of study from the Azure Administrator or Fundamentals exams.
Memorising facts will not get you far. Practise working through scenarios instead:
- Why would someone choose Cosmos DB over SQL Database for a specific workload?
- What are the cost implications of choosing Availability Zones versus Paired Regions?
- How does identity architecture change as an organisation grows from 100 to 10,000 employees?
Read Azure architecture documentation as well as training courses. The Azure Architecture Center documents patterns used by large organisations, and those patterns come up often on the exam.
The azureprep.com AZ-305 practice tests use the case study format and show where your gaps are. Practice tests make you apply what you know under time pressure, as the real exam does, which lectures cannot.
Recommended Study Timeline
For someone with AZ-104 certification and 12+ months of hands-on Azure experience:
- Weeks 1-2: Review AZ-305 domains, watch targeted training content
- Weeks 3-4: Study Domain 1 (identity, governance, monitoring), attempt practice tests
- Weeks 5-6: Study Domain 2 (data storage), complete azureprep.com practice tests
- Weeks 7-8: Study Domains 3 and 4 (business continuity, infrastructure)
- Weeks 9-10: Full practice exams under exam conditions (100 minutes, no breaks)
- Week 11: Review weak areas identified by practice tests
- Week 12: Final practice exam, schedule real exam
Adjust the timeline to your experience and your practice test results.
Exam Tips and Strategies
Read each case study in full before answering. Candidates who skim miss requirements.
Note the details that decide answers as you read: budget, compliance requirements, RTO/RPO targets and geographic spread.
When more than one answer looks correct, choose the one that meets every requirement. AZ-305 often asks why one approach beats alternatives that partly work.
Watch the cost. The hardest questions ask for the solution that meets all requirements at the lowest cost, so you need to know what drives the cost of each Azure service.
You can return to earlier questions. If a question is taking too long, flag it, move on, and come back once the others are answered if time allows.
Look for the practical answer: the one that works reliably at a sensible cost, rather than the theoretically perfect design.
Final Thoughts
The AZ-305 certification shows that you can design Azure solutions for large organisations, and the exam is hard because that work is hard.
Prepare by learning why architects choose particular services for particular problems, as well as what each service does. Use the azureprep.com practice tests to get used to the case study format before exam day.