Azure Fundamentals AZ-900 Sign in Try 10 free questions

Free AZ-900 practice questions: Cloud Concepts

The first domain of Azure Fundamentals at 25–30% of the exam — mostly vocabulary, mostly winnable once the Azure words click. AZ-900, 2026 edition.

Here are 7 free Cloud Concepts questions in the exam's own style, each tied to a real Azure service and explained in plain English — no sign-up.

Straight to the 7 free questions ↓

Cloud Concepts is where almost everyone starts the AZ-900, and there is a reason it feels approachable: it is the most vocabulary-driven area of the exam. You are not being asked to build anything or debug a broken deployment. You are being asked to recognise a situation and give it the right name. That is a very learnable skill, and this page is built to help you practise it with real Azure examples rather than abstract definitions.

Why this domain rewards clear labels

At 25–30% of the exam, Cloud Concepts is the smallest of the three AZ-900 areas — but it sets up everything after it. Most questions here describe a scenario in a sentence or two and ask you to pick the term that fits: is this IaaS or PaaS? Public, private or hybrid? High availability or scalability? If the words are crisp in your head, these questions take seconds. If they blur together, four options all look plausible. So the whole game is turning fuzzy definitions into sharp ones, and the fastest way to do that is to pin each term to a concrete Azure service you can picture.

The service models, seen through Azure

The diagram above is the single most useful picture in this domain. It shows the same stack — applications, data, runtime, operating system, virtualisation, storage and networking — under three service models, shaded by who manages each layer. Read it left to right and the pattern jumps out: the more managed the service, the less of the stack is yours.

  • IaaS — Azure Virtual Machines. You manage everything from the operating system up. Full control, full responsibility, including OS patching.
  • PaaS — Azure App Service. You bring your code and data; Microsoft runs the OS and everything beneath it. You just deploy.
  • SaaS — Microsoft 365. A finished product. You sign in and use it; Microsoft manages the entire stack.

That single line — where the shaded boundary sits — quietly answers a surprising number of exam questions, including who applies a security patch and who is responsible when something below the line breaks.

Deployment models: where the resources live

Separate from the service model is the deployment model, which is about location. Public cloud means your resources live in Microsoft’s data centres. Private cloud means they run on infrastructure dedicated to you, often on-premises. Hybrid cloud combines the two, and the exam loves scenarios where regulation forces some data to stay on-site while the team still wants Azure’s management tools. When you see “on-premises and cloud, managed together,” think hybrid — and think of the Azure services that make it real, like Azure Arc for governing outside servers and Azure Stack for running Azure in your own building.

Concepts that describe behaviour

The last cluster of terms describes how a well-designed Azure workload behaves. High availability keeps a service running through failures — the “is it still up?” question, which you design for with Availability Zones. Scalability is the ability to grow, whether by moving to a bigger machine (scaling up, vertical) or adding more machines (scaling out, horizontal). Elasticity is scaling that happens automatically in both directions, and Azure autoscale on an App Service plan is the textbook example: instances appear when CPU climbs and disappear when the rush passes. Keep these three questions apart — is it up, can it grow, does it adjust itself? — and the distractors stop working on you.

Practise the reasoning, not the answer

Work through the seven questions below one at a time, and read each explanation all the way past the correct option. The “Why not the others:” line is where this domain is actually learned, because on exam day the three wrong answers will look just as tidy as the right one. Get comfortable naming Azure situations here, and Cloud Concepts becomes one of the calmest parts of your test.

A three-column diagram comparing the IaaS, PaaS and SaaS cloud service models by who manages each layer of the stack. Rows from top to bottom are Applications, Data, Runtime and middleware, Operating system, Virtualisation and servers, and Storage and networking. The IaaS column, labelled Azure VMs, is shaded as customer-managed for the top four layers down through the operating system, and Microsoft-managed for virtualisation and storage or networking. The PaaS column, labelled Azure App Service, is customer-managed only for applications and data, with everything from the runtime downward managed by Microsoft. The SaaS column, labelled Microsoft 365, is entirely Microsoft-managed. A caption reads: you patch the OS equals Azure VM, just deploy your code equals App Service, sign in and use it equals Microsoft 365.

7 free AZ-900 practice questions

Answers and explanations — no email wall
AZ-900 Question 1 of 7

Your team needs to run a legacy Windows application on Azure and requires full control of the operating system to install specific drivers and scheduled patches. Which Azure offering matches this need, and which cloud service model is it?

Answer: B — Azure Virtual Machines — Infrastructure as a Service (IaaS).

When the requirement is full control of the operating system, you are describing IaaS, and in Azure that is Azure Virtual Machines — you get the VM and manage everything from the OS up. Why not the others: App Service and Azure Functions are PaaS, where Microsoft owns the OS and you only supply code, so you cannot install your own drivers or control OS patching. Microsoft 365 is a finished SaaS product you sign in to, with no operating system for you to touch at all. Pro tip: the phrase 'control of the operating system' is the exam's tell for IaaS every time.

AZ-900 Question 2 of 7

A hospital must keep patient records on servers in its own on-site data centre for regulatory reasons, but wants to manage and govern those servers using the same Azure tools it uses for its cloud resources. Which deployment model does this describe?

Answer: C — Hybrid cloud, extending Azure management to on-premises servers with Azure Arc.

Keeping some resources on-premises while managing them together with cloud resources is the definition of hybrid cloud, and Azure Arc is the service that projects on-premises (and other) servers into Azure so you govern them with the same tools. Why not the others: a public cloud region would mean the data lives in Microsoft's data centres, which the regulation forbids here. A private cloud with no Azure connection could not use Azure tooling at all, failing the second requirement. Multi-cloud describes using more than one public provider, which is not what the scenario asks for. Pro tip: 'on-premises plus cloud, managed together' equals hybrid — think Azure Arc and Azure Stack.

AZ-900 Question 3 of 7

You run a website on an Azure Virtual Machine and also a separate site on Azure App Service. A critical operating-system security patch is released. Who is responsible for applying that OS patch in each case?

Answer: B — You patch the OS on the Azure VM; Microsoft patches the OS under App Service.

The shared responsibility model shifts with the service model. On an Azure VM (IaaS) the operating system is yours to maintain, so you apply OS patches; on App Service (PaaS) Microsoft manages the underlying OS, so that patch is theirs. Why not the others: 'Microsoft patches both' ignores that IaaS hands you the OS. 'You patch both' wrongly claims the OS under a PaaS service. The last option simply reverses the correct split. Pro tip: the more managed the service, the less of the stack is yours — the diagram above shows exactly where the line moves.

AZ-900 Question 4 of 7

A start-up wants to launch an Azure App Service web app for a weekend product demo and pay only for the compute it actually uses during those two days, with nothing to pay once it shuts the app down. Which Azure pricing characteristic makes this possible?

Answer: B — Consumption-based, pay-as-you-go pricing billed for actual usage.

Paying only for what you consume and nothing once resources stop is consumption-based, pay-as-you-go pricing — a core benefit of Azure that suits short-lived or unpredictable workloads. Why not the others: an annual reservation is a commitment paid ahead of time to lower the rate on steady workloads, the opposite of a two-day demo. A perpetual per-server licence is the on-premises purchasing model cloud replaces. A flat monthly fee would charge you even after you shut the app down. Pro tip: 'pay only for what you use, stop paying when you stop' is the pay-as-you-go answer.

AZ-900 Question 5 of 7

Two requirements land on your desk. Marketing wants your Azure web app to stay online even if one data centre has an outage. Engineering wants it to handle a sudden traffic spike by adding more instances. Which pair of Azure concepts do these describe, in order?

Answer: B — High availability for staying online; scalability for the traffic spike.

Staying online through an outage is high availability — keeping the service running despite failures, which in Azure you design for with Availability Zones. Handling more traffic by adding instances is scalability. Why not the others: the first option swaps the two definitions. Disaster recovery is about recovering after a major regional event, not routine uptime or scaling. Elasticity and reliability are related ideas, but they do not match this specific pairing. Pro tip: high availability answers 'is it up?', scalability answers 'can it grow?' — keep those two questions separate.

AZ-900 Question 6 of 7

Your Azure App Service plan is set with autoscale rules that add instances when average CPU rises above 70% and remove them when it falls below 30%. During a marketing campaign, traffic surges for three hours and then returns to normal. What benefit of the cloud does this behaviour illustrate?

Answer: A — Elasticity — capacity automatically grows and shrinks to match demand.

Automatically adding capacity when demand rises and releasing it when demand falls is elasticity, and Azure autoscale is exactly how App Service delivers it. Why not the others: fault tolerance is about surviving failures, not matching capacity to load. Data residency concerns where data is physically stored, unrelated to scaling. Vendor lock-in is a trade-off consideration, not a benefit or a scaling behaviour. Pro tip: elasticity is scaling that happens on its own, up and back down — autoscale is the Azure feature that proves it.

AZ-900 Question 7 of 7

Your Azure Virtual Machine is running out of memory under load. You could move it to a larger VM size with more RAM, or you could keep the size and add more identical VMs behind a load balancer. Which options are these, in order?

Answer: B — Scaling up (vertical), then scaling out (horizontal).

Moving to a larger VM size adds power to a single machine — that is scaling up, or vertical scaling. Adding more identical VMs behind a load balancer spreads load across many machines — that is scaling out, or horizontal scaling. Why not the others: the first option reverses the two. Claiming both are the same type ignores that one changes a machine's size while the other changes the machine count. Pro tip: 'up' equals bigger box, 'out' equals more boxes — Azure Virtual Machine Scale Sets are the classic scale-out example.

That is exactly how every question in the course works — answer, explanation, why-not. The real set continues in the practice player: 10 free questions, no sign-up.

Continue with the 10 free questions →

Those 7 questions were the start.

The exam does not test whether you recognise a term — it tests whether you can rule out three plausible answers under time pressure. That is what the explanations above are for, and there are 300 more questions built exactly like them.

Collecting questions yourself

  • Scattered across forums, of unknown age
  • Answer keys without reasoning
  • No idea which domain you are weak in

Practising with a system

  • 300 questions in 6 full tests, AZ-900 (2026)
  • Every option explained — including the wrong ones
  • Readiness per exam domain, and drills for your weakest
Start free — 10 questions

Straight into the player. No account, no email.

Frequently asked questions

How much of the AZ-900 exam is Cloud Concepts?
Microsoft's official exam skills outline weights Cloud Concepts at roughly 25–30% of the exam — the smallest of the three areas, alongside Azure Architecture and Services (35–40%) and Azure Management and Governance (30–35%). It is a stated range, not a fixed number, so treat it as a band. Our simulator draws each practice run in these official blueprint proportions, so a full-length test gives Cloud Concepts the same share of your time the real exam will.
Why do so many Cloud Concepts questions feel like vocabulary tests?
Because they mostly are. This domain checks whether you can attach the right label — IaaS, PaaS, SaaS, hybrid, elasticity, high availability — to a described situation. The good news is that once the definitions click, these become some of the fastest points on the exam. Anchor each term to one concrete Azure service, like Azure VMs for IaaS or App Service for PaaS, and the labels stop blurring together.
Are these real exam questions?
No. Every question here is written by us in the style and difficulty of the exam, never copied from any question bank or the real test. Sharing real exam items breaks Microsoft's certification agreement and can cost a candidate their certification — and memorising an item teaches you nothing about the one you have not seen. Learning the reasoning behind each answer is what actually transfers to exam day.

Updated for AZ-900 (July 2026). The sample questions above are our own work in the style of the exam — not real exam items. The exam itself is set and marked by the certification body.