Azure Fundamentals (AZ-900) — All Questions

Which statement best defines a consumption-based (pay-as-you-go) pricing model?

• a.You must estimate and pay for peak capacity even when it sits completely idle
• b.You pay only for the resources you use, with no upfront commitment✓
• c.A fixed yearly fee is paid in advance regardless of how much you actually use
• d.Billing is based on the number of employees in the organization

Consumption-based pricing charges you only for the compute, storage, and services you actually consume. There is no large upfront commitment, so spending tracks real usage. This lets organizations avoid paying for idle, over-provisioned capacity.

Which spending shift best characterizes moving from an on-premises datacenter to a pay-as-you-go cloud model?

• a.From variable usage-based costs to fixed hardware purchases
• b.From subscription billing to one-time perpetual software licenses
• c.From capital expenditure (CapEx) to operating expenditure (OpEx)✓
• d.From operating expenditure (OpEx) to capital expenditure (CapEx)

On-premises IT relies on CapEx: large upfront purchases of servers and facilities. The cloud shifts spending to OpEx, an ongoing usage-based cost. This improves cash flow and removes the need to buy assets years ahead of demand.

Which of the following is the clearest example of Software as a Service (SaaS)?

• a.A managed database platform where you still deploy your own schema and application code
• b.A bare-metal server hosted in a colocation facility that you fully manage
• c.A ready-to-use, web-based email service accessed through a browser✓
• d.Renting a virtual machine on which you install and maintain your own applications yourself

SaaS delivers a finished application that users simply consume, usually through a browser. The provider manages everything from hardware up to the application. Hosted email or CRM products are classic SaaS examples.

A company wants full control of the operating system and installed software but does not want to buy or maintain physical servers. Which model fits?

• a.Platform as a Service (PaaS)
• b.Infrastructure as a Service (IaaS)✓
• c.Software as a Service (SaaS)
• d.A fully managed, ready-to-use application delivered over the internet

IaaS provides virtualized compute, storage, and networking while the provider maintains the physical hardware. The customer controls the OS and everything above it. This offers maximum flexibility without owning servers.

In the cloud shared responsibility model, which item is ALWAYS the provider's responsibility across every service model?

• a.The configuration of user access permissions
• b.The classification and protection of the customer's data
• c.The management of user accounts and identities
• d.The physical security of the datacenters and hardware✓

No matter the service model, the provider always secures the physical datacenter and hardware. Data, identities, and access remain the customer's responsibility. Layers in between shift depending on IaaS, PaaS, or SaaS.

How does consumption-based pricing help organizations avoid wasted spending?

• a.You pay only for what you use, avoiding over-provisioning for peak demand✓
• b.It charges a flat fee that never changes regardless of how much you use
• c.It bills based on the number of datacenters the provider operates worldwide
• d.It requires you to buy three years of capacity in advance for every workload

Because you pay only for what you consume, you no longer buy hardware sized for peak load that sits idle most of the time. Costs scale up and down with demand. This eliminates the waste of traditional over-provisioning.

Which statement is TRUE of the public cloud deployment model?

• a.It can only be used by a single organization inside its own building
• b.A third-party provider owns the infrastructure and offers it to many customers✓
• c.The customer owns, maintains, and refreshes all of the underlying physical hardware
• d.It cannot scale beyond a small fixed amount of capacity

In a public cloud, a third-party provider owns and operates the infrastructure and offers it to many tenants over the internet. Resources are shared and delivered on demand. This gives low upfront cost and vast scale.

An organization needs computing resources used exclusively by itself for compliance reasons, hosted either on-premises or by a provider. Which model is this?

• a.Community cloud shared by several agencies
• b.Private cloud✓
• c.Multi-tenant SaaS
• d.Public cloud

A private cloud dedicates infrastructure to a single organization, whether hosted internally or by a provider. This isolation supports strict security and compliance needs. The trade-off is usually higher cost and management effort.

What is a key advantage of a hybrid cloud approach?

• a.It forces all data to remain permanently in a single datacenter
• b.It eliminates the ability to use any public cloud services
• c.It lets sensitive workloads stay on-premises while bursting to the public cloud for scale✓
• d.It completely removes the need for any form of internet connectivity between sites

Hybrid cloud links private, on-premises environments with public cloud services. Sensitive systems can stay local while workloads burst to the cloud for extra capacity. It suits gradual migration and regulatory requirements.

Which description best matches elasticity rather than plain scalability?

• a.Storage is limited to one physical disk per server
• b.Resources automatically grow and shrink in near real time as demand changes✓
• c.An administrator manually places an order for additional servers whenever demand rises
• d.Capacity is fixed at deployment and never changes afterward

Elasticity is the automatic, demand-driven expansion and contraction of resources. Scalability is the broader ability to grow capacity, which can be manual. Elasticity specifically implies near real-time, hands-off adjustment.

Which design choice most directly increases the high availability of an application?

• a.Running the application on a single server with no backups at all
• b.Storing every copy of the data on one disk in one datacenter
• c.Deploying redundant instances across multiple fault domains✓
• d.Disabling monitoring and health checks to reduce operating cost

High availability comes from redundancy so that no single failure takes the service down. Spreading instances across fault domains keeps the app running during hardware faults. Monitoring and backups support, but do not by themselves create, availability.

A startup can launch a new test environment in minutes instead of waiting weeks for hardware. Which cloud benefit does this illustrate?

• a.Multi-tenancy
• b.Agility✓
• c.Data sovereignty
• d.Vendor lock-in

Agility is the ability to provision resources and deliver solutions quickly. Instead of long hardware procurement, environments spin up in minutes. This speeds experimentation and time to market.

Which practice best supports disaster recovery for a cloud-hosted application?

• a.Turning off all backups permanently to reduce monthly cost
• b.Replicating data and workloads to a geographically separate region✓
• c.Relying on users to email themselves copies of important files
• d.Keeping only one single copy of all production data in one primary region

Disaster recovery restores systems and data after a major outage. Replicating to a separate region protects against a whole-region failure such as a natural disaster. This minimizes downtime and data loss.

Which of the following is an example of a capital expenditure (CapEx)?

• a.A one-time purchase of physical servers for a datacenter✓
• b.A per-second charge for a running container instance
• c.A usage-based fee for outbound data transfer
• d.A recurring monthly invoice for the cloud compute hours actually consumed

CapEx is an upfront purchase of long-lived assets such as servers or datacenter equipment. It is depreciated over years. The cloud reduces CapEx by replacing it with usage-based OpEx.

Which characteristic describes operating expenditure (OpEx) in a cloud context?

• a.Ongoing, usage-based costs that scale with consumption✓
• b.A large upfront purchase of hardware assets
• c.A fixed asset that is depreciated over several years
• d.A one-time perpetual software license bought outright

OpEx is ongoing spending that rises and falls with usage, like a monthly cloud bill. It contrasts with CapEx, a big upfront asset purchase. The pay-as-you-go cloud model favors OpEx.

Under PaaS, which responsibility typically remains with the customer?

• a.Managing their application code and data✓
• b.Maintaining the physical network hardware in the datacenter
• c.Patching the underlying host operating system
• d.Replacing failed physical disks in the datacenter

With PaaS the provider manages the OS, runtime, and infrastructure. The customer focuses on their application code and data. This reduces operational overhead while keeping control of the app.

Adding more identical web server instances behind a load balancer to handle more users is an example of what?

• a.Archiving cold data to lower tiers
• b.Horizontal scaling (scaling out)✓
• c.Reducing redundancy across the system
• d.Vertical scaling (scaling up)

Horizontal scaling, or scaling out, adds more instances to share the load. Vertical scaling instead adds CPU or memory to a single machine. Scaling out is often more resilient and elastic.

How does fault tolerance differ from disaster recovery?

• a.Disaster recovery prevents any failure from ever happening at all
• b.Fault tolerance only applies to monthly billing calculations
• c.Fault tolerance keeps a system running through component failures; disaster recovery restores it after a major outage✓
• d.They are exactly the same concept and are fully interchangeable in every single situation

Fault tolerance uses redundancy so individual component failures do not stop the service. Disaster recovery is about restoring service and data after a large disruptive event. They are complementary reliability strategies.

If a cloud service offers a higher composite SLA, what does that generally indicate?

• a.It consumes less electricity in the datacenter
• b.It offers a stronger guarantee of uptime and availability✓
• c.It costs nothing at all to run each month
• d.It is far more likely to contain unpatched security vulnerabilities over time

An SLA is the provider's formal commitment to a level of uptime and availability. A higher SLA means a stronger guarantee and less allowed downtime. Failing to meet it can entitle customers to service credits.

Why does multi-tenancy help lower costs in the public cloud?

• a.Because it removes the need for any security controls whatsoever
• b.Because each and every customer is given fully dedicated physical hardware of their own
• c.Because it stores all data on the customer's own premises
• d.Because many customers share the same underlying infrastructure efficiently✓

Multi-tenancy lets many customers share pooled physical resources while staying logically isolated. This high utilization spreads cost across tenants. Providers enforce strong isolation to keep tenants secure from one another.

Cloud providers can often offer lower per-unit prices than a single company running its own datacenter. Why?

• a.Because they never invest in any new hardware at all
• b.Economies of scale from operating massive, shared datacenters✓
• c.Because customers must pay for all unused capacity upfront
• d.Because they avoid every electricity and cooling cost entirely

Providers operate at enormous scale, buying and running hardware far more efficiently than a single firm. These economies of scale lower the per-unit cost. Savings are passed on through competitive pricing.

A team wants to build and deploy a web app while letting the provider handle OS patching, runtime, and scaling. Which model should they choose?

• a.Infrastructure as a Service (IaaS)
• b.On-premises bare metal
• c.Colocation hosting
• d.Platform as a Service (PaaS)✓

PaaS provides a managed platform including OS, runtime, and scaling. Developers deploy code without managing servers. This accelerates delivery and cuts operational work.

With a SaaS product, what does the customer still control?

• a.The physical servers running the service
• b.The operating system patching schedule and the maintenance windows for the servers
• c.The application's underlying source code
• d.Their own data and which users have accounts✓

In SaaS the provider manages almost the entire stack. The customer's remaining responsibilities center on their data, users, and access settings. This makes SaaS the lowest-maintenance model for consumers.

Which billing behavior is characteristic of many serverless and consumption-based cloud services?

• a.A single flat annual license fee that applies regardless of how much you actually use
• b.Charges that accrue per execution or per second of actual use✓
• c.Charges that occur only when you buy new physical hardware
• d.A single fixed fee for truly unlimited usage forever

Serverless and consumption services bill on granular actual usage, such as per execution or per second. There is no charge when nothing runs. This tightly aligns cost with real demand.

A hospital must keep certain patient records on infrastructure it fully controls but wants cloud scale for its public website. Which model fits both needs?

• a.SaaS only
• b.Public cloud only
• c.Hybrid cloud✓
• d.A single private cloud with no internet access at all

A hybrid cloud lets the hospital keep sensitive records on controlled infrastructure while using public cloud scale for the website. Data and apps can move between the two. This balances compliance with flexibility.

How does deploying to multiple global regions improve the user experience?

• a.It increases the distance data must travel to users
• b.It removes the need for any redundancy in the design
• c.It forces all traffic through one single country
• d.It places resources closer to users, lowering latency✓

Running resources in regions near users reduces network latency. It also supports resilience and data-residency compliance. Global reach is a core advantage of large cloud platforms.

In a traditional on-premises datacenter (not cloud), who is responsible for the physical security of the servers?

• a.No one is responsible for it in that model
• b.The external cloud provider that hosts the organization's other workloads
• c.The internet service provider that supplies connectivity
• d.The organization that owns the datacenter✓

On-premises, the owning organization is fully responsible for physical security, hardware, and everything else. Moving to the cloud shifts physical security to the provider. This is a key difference in the shared responsibility model.

During off-peak hours, an elastic system automatically removes unused instances. What is the main benefit?

• a.Lower cost by not paying for idle capacity✓
• b.A guarantee that the customer owns the physical hardware
• c.Permanently fixed capacity that never changes
• d.Consistently higher latency and slower response times for all end users

Elasticity releases resources when demand falls so you stop paying for idle capacity. Capacity returns automatically when demand rises. This keeps performance steady while controlling cost.

Which cloud capability most improves reliability during a hardware failure?

• a.Turning off all health monitoring to save money
• b.Deleting older backups on a regular schedule to conserve storage space
• c.Automatic failover to redundant resources✓
• d.Using only a single availability zone for all resources

Reliability during failures depends on redundancy and automatic failover. When one component fails, traffic shifts to healthy resources. This keeps the service available with minimal disruption.

What does 'scalability' primarily describe?

• a.The ability to increase capacity to handle growing workloads✓
• b.The ability to encrypt data both at rest and in transit automatically
• c.The ability to physically own the underlying servers
• d.The ability to disable all logging across the system

Scalability is the capacity to grow resources to meet increasing demand. It can be achieved by scaling up or scaling out. It ensures applications keep performing as workloads expand.

Why might a growing startup prefer the cloud's OpEx model over large CapEx purchases?

• a.It locks them into rigid ten-year hardware purchase and maintenance contracts
• b.It requires buying a full datacenter building first before any deployment
• c.It preserves cash by avoiding large upfront hardware investments✓
• d.It guarantees the software they build will contain no bugs

The OpEx model spreads cost over time as usage-based charges. This preserves cash and avoids risky upfront hardware bets. Startups can scale spending gradually as they grow.

You deploy a virtual machine using IaaS. Who applies security updates to the guest operating system?

• a.The hardware manufacturer handles it automatically
• b.The provider does it silently on your behalf
• c.No operating system updates are ever needed in IaaS
• d.You, the customer✓

With IaaS the customer manages the guest OS, including patching. The provider only maintains the physical host and hypervisor. This division is central to cloud security responsibilities.

An enterprise gradually moving workloads to the cloud over several years, keeping some systems on-premises meanwhile, is using which approach?

• a.A hybrid cloud migration✓
• b.Colocation hosting only
• c.SaaS used exclusively for everything
• d.Pure public cloud adopted from day one

Running some systems on-premises while others move to the cloud is a hybrid approach. It supports phased migration and reduces risk. Workloads can be shifted as they are ready.

In disaster recovery planning, what does the Recovery Time Objective (RTO) represent?

• a.The target time to restore a system after an outage✓
• b.The total number of users on the system
• c.The total amount of storage the system uses
• d.The total monthly cost of the service including data transfer charges

RTO is the maximum acceptable time to restore service after a disruption. It guides how fast recovery mechanisms must work. A related metric, RPO, defines acceptable data loss.

Compared with pay-as-you-go, what do you trade for a discount when you reserve capacity for one or three years?

• a.You automatically lose all technical support entitlements
• b.You must immediately move to a competing cloud provider
• c.You commit to a term in exchange for lower rates✓
• d.You receive noticeably worse performance for the same reserved workloads

Reservations give a discount in return for committing to a one- or three-year term. They suit steady, predictable workloads. The trade-off is reduced flexibility, not reduced performance.

Which of the following is NOT typically a benefit of public cloud computing?

• a.Reduced upfront capital cost through a pay-as-you-go consumption model
• b.Global reach across many regions
• c.On-demand scalability of resources
• d.Full physical control over the underlying hardware✓

Public cloud customers give up physical control of hardware to the provider. The recognized benefits include scalability, lower upfront cost, and global reach. Trading physical control for these advantages defines the model.

Regardless of using IaaS, PaaS, or SaaS, which security duty stays with the customer?

• a.Securing the physical building and its entrances
• b.Cooling and powering the physical datacenter
• c.Replacing failed network switches and other hardware in the server racks
• d.Managing their own user identities and access✓

Across every model the customer remains responsible for their data, identities, and access. The provider always handles the physical facility and hardware. Understanding this boundary is essential to cloud security.

An online retailer's site automatically adds capacity on a busy shopping holiday and releases it afterward. This is a direct example of what?

• a.Fixed, static provisioning of resources
• b.Data residency requirements
• c.Elasticity✓
• d.Vendor lock-in to a single provider

Automatically adding and removing capacity in response to demand is elasticity. It keeps the site fast during spikes and cheap during lulls. This is a hallmark of well-designed cloud systems.

What is a common trade-off of choosing a private cloud over the public cloud?

• a.It cannot enforce any security controls at all
• b.It usually costs more and requires more management effort✓
• c.It always provides completely unlimited capacity at no additional cost whatsoever
• d.It removes the ability to control who has access

Private clouds offer greater control and isolation but typically cost more and demand more management. Capacity is limited by what the organization provisions. The benefit is control; the cost is effort and expense.

How does cloud agility support innovation?

• a.By locking teams into one fixed design permanently
• b.By letting teams quickly spin up and tear down environments to experiment✓
• c.By actively preventing any experimentation by teams
• d.By requiring long hardware procurement cycles before any new idea can be tested

Agility lets teams create and discard environments in minutes. Cheap, fast experimentation lowers the cost of trying new ideas. This accelerates innovation and learning.

How do high availability and disaster recovery differ?

• a.High availability only relates to monthly billing
• b.They are two identical concepts that mean exactly the same thing in every practical case
• c.High availability minimizes downtime day to day; disaster recovery restores service after a major event✓
• d.Disaster recovery guarantees that downtime can never occur

High availability keeps services running through routine failures with minimal downtime. Disaster recovery focuses on restoring service after a major, often regional, disruption. Robust designs use both.

When comparing on-premises to cloud, which cost is often overlooked but included in a true total cost of ownership?

• a.The particular color and finish chosen for the server chassis
• b.Power, cooling, and datacenter staffing✓
• c.The brand of coffee in the office kitchen
• d.The number of company holidays each year

Total cost of ownership includes hidden operational costs like electricity, cooling, and staff. These are easy to overlook when comparing to cloud pricing. Accounting for them gives a fairer comparison.

What is one risk of the consumption-based model if usage is not monitored?

• a.Physical theft of hardware directly from the provider's secured datacenter
• b.Unexpectedly high bills from unmonitored usage✓
• c.Guaranteed perfectly fixed bills every month
• d.A complete inability for costs to change at all

Because charges scale with usage, unmonitored growth can lead to surprise bills. Cost controls, budgets, and alerts help prevent this. Monitoring spending is a key cloud discipline.

Several government agencies with shared compliance needs pool infrastructure used only among themselves. This is best described as what?

• a.A pure public cloud open to everyone
• b.A SaaS application
• c.A community cloud✓
• d.A strictly single-tenant private cloud for one agency

A community cloud is shared by several organizations with common requirements, such as compliance. It sits between private and public models. Costs and governance are shared among the members.

What distinguishes autoscaling from manual scaling?

• a.Autoscaling requires placing a phone call to the provider's support team each time
• b.Manual scaling needs no human action whatsoever
• c.Autoscaling adjusts capacity automatically based on defined rules✓
• d.Autoscaling can only reduce capacity and never increase it

Autoscaling changes capacity automatically according to metrics or schedules. Manual scaling requires a person to act. Autoscaling supports elasticity with minimal operational effort.

Automated, regular backups in the cloud primarily support which goal?

• a.Data durability and recoverability✓
• b.Reduced font sizes in applications
• c.Faster graphics rendering for video editing workloads on the desktop
• d.Lower physical screen brightness on monitors

Regular backups protect against data loss and enable recovery after failures or mistakes. They underpin reliability and disaster recovery. Cloud platforms make backups easy to automate.

Shifting IT spending from CapEx to OpEx can improve which business measure?

• a.The total square footage of the company's physical office building space
• b.Cash flow and financial flexibility✓
• c.The design of the company logo
• d.The number of parking spaces available to staff

Replacing large upfront purchases with pay-as-you-go spending improves cash flow. Money is not tied up in depreciating hardware. This flexibility helps organizations adapt to change.

As you move from IaaS to PaaS to SaaS, how does the customer's share of responsibility generally change?

• a.It increases steadily at each step
• b.It only becomes the provider's responsibility after a continuous ten-year period
• c.It decreases as more is managed by the provider✓
• d.It stays exactly the same across all three models

Moving from IaaS toward SaaS, the provider manages more of the stack. The customer's responsibilities shrink accordingly. SaaS leaves the customer mainly responsible for data and access.

Which term specifically implies automatic, demand-driven adjustment rather than a one-time capacity increase?

• a.Colocation of servers
• b.Vertical scaling of a single machine
• c.Capital expenditure on new hardware
• d.Elasticity✓

Elasticity means resources automatically expand and contract with demand. A one-time capacity increase is scaling, not necessarily elastic. Elasticity is the automatic, real-time form.

A common way to describe the agility benefit of cloud is improved…

• a.datacenter floor tiling and paint
• b.time to market for new solutions✓
• c.network cable color coding standards
• d.physical server weight and the total rack loading in the facility

Agility shortens the time from idea to deployed solution. Fast provisioning removes long hardware delays. This improved time to market is a central cloud advantage.