AZ-305 Explained: What Designing Azure Infrastructure Solutions Really Tests
AZ-305 doesn't test what you know about Azure services — it tests how you decide between them. Here's what the exam actually measures.

AZ-305 is not really an Azure knowledge test — it's a decision-making test that happens to be set in Azure. Officially titled Designing Microsoft Azure Infrastructure Solutions, this exam is the core requirement for the Microsoft Certified: Azure Solutions Architect Expert credential, and it has a well-earned reputation as one of the trickier exams in the Microsoft cloud track. It's not because the services are exotic. It's because almost every question hands you a business constraint — a budget, a compliance requirement, a latency target, an existing on-prem investment — and asks you to pick the design that actually fits, not just a design that would technically work. Candidates who know Azure well but haven't practiced thinking like an architect are often the ones who walk out surprised by their score.
What AZ-305 actually measures
Microsoft organizes the exam into four domains, and the weighting tells you a lot about where to spend your study time. Designing identity, governance, and monitoring solutions makes up roughly a quarter to a third of the exam and covers Microsoft Entra ID, role-based access control, Azure Policy, management group hierarchies, and observability with Azure Monitor and Log Analytics. Designing data storage solutions is next, covering how to choose between relational and non-relational stores, when Azure SQL beats Cosmos DB, how Blob Storage tiers and redundancy options fit different retention needs, and how data actually moves between systems securely. Designing business continuity solutions focuses on backup strategy, disaster recovery architecture, and high availability — the stuff that matters when something breaks. And the largest domain, designing infrastructure solutions, covers compute choices across VMs, containers, and serverless; application architecture patterns like messaging, event-driven design, and API management; migration strategy using the Cloud Adoption Framework and Azure Migrate; and network design covering connectivity, performance, and security.
In practice, expect somewhere around 40 to 60 questions in a 120-minute window, a passing score of 700 out of 1000, and a cost in the neighborhood of $165 USD — though you should always confirm current pricing and format on Microsoft's own exam page before you register, since these details do shift.
Why this exam assumes you already know AZ-104
Microsoft doesn't enforce a hard prerequisite for AZ-305, but the exam is written as if you've already passed AZ-104, Microsoft Azure Administrator, or have equivalent hands-on experience. That distinction matters more than most candidates realize going in. AZ-104 tests whether you can configure a virtual network, deploy a VM scale set, or set up a backup vault. AZ-305 assumes you already know how to do those things and instead asks whether you should — and under what conditions you'd choose a different approach entirely. If you've never actually built anything in Azure, you'll struggle here regardless of how many practice questions you memorize, because the exam is testing judgment that's built from experience, not just terminology recall.
The trade-off questions that trip people up
The single most commonly confused concept on AZ-305 is when to use Azure SQL Database versus Azure SQL Managed Instance versus SQL Server on a VM. Candidates often default to whichever option they're most familiar with, but the exam is testing whether you understand the trade-off: Managed Instance exists specifically to bridge the gap when an application needs SQL Server Agent, cross-database queries, or linked servers that Azure SQL Database's PaaS model doesn't support, while a VM is the answer only when you need OS-level control or a legacy feature nothing else supports. Get comfortable articulating why, not just which.
A second frequent stumbling block is choosing between Azure Firewall, Network Security Groups, and Application Gateway with a Web Application Firewall. These aren't interchangeable, and scenario questions love to describe a threat (say, SQL injection against a web app) that only one of them actually addresses at the right layer. A third is disaster recovery versus high availability — candidates conflate the two constantly, but HA is about surviving a failure within a region with minimal downtime, while DR is about recovering in a different region after a larger-scale outage, and the recovery time and recovery point objectives in the scenario tell you which one is actually being asked about. Read every case study for its RTO and RPO numbers before you touch the answer choices; they're often the whole question in disguise.
How to actually prepare, not just read about it
Reading Microsoft Learn modules will get you vocabulary, but AZ-305 rewards pattern recognition built from repetition across realistic scenarios — noticing that "the company wants to retain compliance data for seven years at the lowest cost" is a signal for Archive tier storage almost every time, or that "minimize downtime during a regional outage with near-zero data loss" is pointing you toward synchronous replication and an active-active design, not a nightly backup. That kind of instinct doesn't come from re-reading documentation; it comes from working through enough scenario-style questions that the patterns start to feel familiar before you finish reading the prompt.
This is exactly where deliberate practice pays off more than passive review. Working through AZ-305 practice questions that mirror the case-study format forces you to justify a design choice the way the real exam does, rather than just recognizing a service name. ExamStudyApp's adaptive practice tracks which of the four domains you're weakest in — say, business continuity or identity governance — and keeps surfacing questions from that area until it stops being a weak spot, instead of having you burn time re-answering questions on compute design you already have down cold.
Knowing when you're actually ready
Because AZ-305 is scenario-heavy, a good gut check for readiness is whether you can read a case study, identify the two or three genuinely competing options, and explain out loud why you're eliminating the wrong ones — not just land on the right one by pattern-matching a keyword. Once you can do that consistently across all four domains, it's worth running a full timed mock exam that mirrors the real 120-minute format and the 700-point passing bar. Timed simulation matters here specifically because AZ-305 case studies are long to read, and candidates who've only ever practiced untimed often discover on exam day that reading comprehension under a clock is its own skill.
ExamStudyApp's readiness tracking exists for exactly this moment — instead of guessing whether you're prepared based on how confident you feel, you get a clearer signal built from your actual performance across every domain and objective. And when you do miss a question, the mistake review attaches an explanation you can actually learn from, which matters enormously on an exam like this where the wrong answer is often "technically works but isn't the best design" rather than flatly incorrect.
The bottom line
AZ-305 rewards architects who can weigh cost, compliance, performance, and operational overhead against each other and defend a choice — not candidates who've simply memorized what each Azure service does. If you've got real hands-on Azure experience or an AZ-104 behind you, the path forward is deliberate, scenario-based practice across identity and governance, data storage, business continuity, and infrastructure design, followed by a realistic timed run before you book the real thing. Designing Microsoft Azure Infrastructure Solutions is a demanding exam, but it's a fair one: it tests the exact judgment you'll need on the job the day after you pass.


