Pass Microsoft Azure AZ-900 Exam in First Attempt Easily
Real Microsoft Azure AZ-900 Exam Questions, Accurate & Verified Answers As Experienced in the Actual Test!

Verified by experts
3 products

You save $69.98

AZ-900 Premium Bundle

  • Premium File 482 Questions & Answers
  • Last Update: Nov 4, 2026
  • Training Course 85 Lectures
  • Study Guide 425 Pages
$79.99 $149.97

Purchase Individually

  • Premium File

    482 Questions & Answers
    Last Update: Nov 4, 2026

    $76.99
    $69.99
  • Training Course

    85 Lectures

    $43.99
    $39.99
  • Study Guide

    425 Pages

    $43.99
    $39.99

Microsoft Azure Fundamentals AZ-900 Practice Test 20

Question 190

A web application must be reachable even if one VM stops responding unexpectedly.

Which cloud or Azure approach best meets the stated requirement?

  1. Use a load balancer plus independently healthy backend instances to remove a single-VM serving dependency.
  2. Deploy suitable zone-redundant resources or instances across availability zones.
  3. Use availability design that distributes instances across appropriate host update and failure domains.
  4. Apply horizontal scaling by increasing the number of suitable service instances.

Explanation

Correct Answer

A:

Load balancing distributes traffic to eligible healthy endpoints; a standalone server remains a single point of failure. This is the implementation most closely tied to the actual requirement.

Incorrect Answers

B:

Availability zones separate local failure domains, reducing exposure to a single datacenter failure when application design supports it.

C:

High availability requires multiple independent serving paths; placing everything on one host defeats redundancy.

D:

Scale-out adds instances to handle load, unlike scale-up which enlarges an individual machine.

Question 191

A development group wants to build a custom business web application while a vendor takes care of runtime scaling and maintenance.

Which observed result would best demonstrate that the intended requirement has been met?

  1. The guest OS can be configured and updated by the tenant while the physical host remains abstracted.
  2. Users complete the required workflows in the hosted product without an internal application delivery team.
  3. Only a generic dashboard status is checked, with no measurement of the specific requirement.
  4. Code releases are deployed through platform tools while OS and runtime servicing remain managed by the service.

Explanation

Correct Answer

D:

PaaS allows coding and deployment while delegating much of server operation to the provider.

Incorrect Answers

A:

That observation could confirm that the guest OS can be configured and updated by the tenant while the physical host remains abstracted, but it does not establish that code releases are deployed through platform tools while OS and runtime servicing remain managed by the service.

B:

That observation could confirm that users complete the required workflows in the hosted product without an internal application delivery team, but it does not establish that code releases are deployed through platform tools while OS and runtime servicing remain managed by the service.

C:

A broad green status cannot substitute for checking the stated outcome. Code releases are deployed through platform tools while OS and runtime servicing remain managed by the service.

Question 192

A team leaves its VM running overnight at zero CPU utilization and expects Azure to bill nothing.

Which observed result would best demonstrate that the intended requirement has been met?

  1. The following billing period reflects fewer resource-hours or stored gigabytes after approved cleanup.
  2. The chargeback report shows elevated spend during the test surge and lower spend after scale-in.
  3. A billing comparison shows charges for a running idle VM and reduced compute charges after deallocation.
  4. The bill follows recorded resource use and resources stop incurring compute charges when properly deallocated.

Explanation

Correct Answer

C:

Metered cloud billing depends on pricing dimensions such as runtime, size and storage allocation, not whether a human is typing.

Incorrect Answers

A:

That observation could confirm that the following billing period reflects fewer resource-hours or stored gigabytes after approved cleanup, but it does not establish that a billing comparison shows charges for a running idle VM and reduced compute charges after deallocation.

B:

That observation could confirm that the chargeback report shows elevated spend during the test surge and lower spend after scale-in, but it does not establish that a billing comparison shows charges for a running idle VM and reduced compute charges after deallocation.

D:

That observation could confirm that the bill follows recorded resource use and resources stop incurring compute charges when properly deallocated, but it does not establish that a billing comparison shows charges for a running idle VM and reduced compute charges after deallocation.

Question 193

A team uses event-driven compute for sending emails but assumes any function invocation runs exactly once.

Which cloud or Azure approach best meets the stated requirement?

  1. Explain that serverless uses provider-managed servers hidden from application operators.
  2. Make serverless event handlers idempotent and design for possible retries.
  3. Assume all available cloud products are operationally interchangeable without checking the workload constraints.
  4. Retain the current manual process as a permanent answer while leaving the documented service requirement unmet.

Explanation

Correct Answer

B:

Event-driven systems can redeliver messages; serverless infrastructure management does not guarantee exactly-once business effects. This is the implementation most closely tied to the actual requirement.

Incorrect Answers

A:

Serverless removes the need to manage servers, not the existence of underlying computing hardware.

C:

Different cloud capabilities have different responsibilities, limits and connectivity behavior. Event-driven systems can redeliver messages; serverless infrastructure management does not guarantee exactly-once business effects.

D:

Continuing the deficient manual process leaves the stated need unresolved. Event-driven systems can redeliver messages; serverless infrastructure management does not guarantee exactly-once business effects.

Question 194

An IT group wants to see current service health and manage multiple cloud applications from one interface.

Which cloud or Azure approach best meets the stated requirement?

  1. Use provider management and monitoring services for centralized administration and visibility.
  2. Use provider services with redundancy and recovery design to improve reliability.
  3. Create least-privilege management roles and governance ownership for cloud changes.
  4. Set measurable performance goals and select suitable geographic placement and architecture.

Explanation

Correct Answer

A:

Cloud manageability includes API/portal-based deployment, configuration, monitoring and automation for distributed services. This is the implementation most closely tied to the actual requirement.

Incorrect Answers

B:

Reliability means delivering intended service over time despite failures; it depends on architecture as well as provider infrastructure.

C:

Cloud governance is an organizational control over who can do what and which rules apply, not a property that appears automatically on cloud adoption.

D:

Provider-managed services offer capabilities, not universal application latency guarantees; network distance and topology still matter.

Question 195

A CFO reviews the financial treatment of a one-time server purchase versus monthly cloud consumption.

Which observed result would best demonstrate that the intended requirement has been met?

  1. The bill follows recorded resource use and resources stop incurring compute charges when properly deallocated.
  2. Cost reports reconcile actual consumption against estimates and identify dominant metered resources.
  3. The estimate identifies resource-specific usage quantities and units rather than one generic price per user.
  4. A cost model separates upfront assets from service bills and applies the correct assumptions for each.

Explanation

Correct Answer

D:

CAPEX typically buys durable assets; OPEX is recurring service consumption, though accounting treatment depends on local rules.

Incorrect Answers

A:

That observation could confirm that the bill follows recorded resource use and resources stop incurring compute charges when properly deallocated, but it does not establish that a cost model separates upfront assets from service bills and applies the correct assumptions for each.

B:

That observation could confirm that cost reports reconcile actual consumption against estimates and identify dominant metered resources, but it does not establish that a cost model separates upfront assets from service bills and applies the correct assumptions for each.

C:

That observation could confirm that the estimate identifies resource-specific usage quantities and units rather than one generic price per user, but it does not establish that a cost model separates upfront assets from service bills and applies the correct assumptions for each.

Question 196

A manufacturer must keep factory controllers locally while exposing analytics dashboards from Azure.

Which cloud or Azure approach best meets the stated requirement?

  1. Choose the public-cloud deployment model with appropriate tenant and access controls.
  2. Distinguish multicloud across providers from hybrid deployments combining private/on-premises and public clouds.
  3. Use a hybrid-cloud approach that connects retained local systems with public-cloud services.
  4. Choose a private-cloud model for the strictly controlled dedicated environment.

Explanation

Correct Answer

C:

Hybrid cloud integrates on-premises or private resources with public cloud so different workloads can reside where their constraints are met. This is the implementation most closely tied to the actual requirement.

Incorrect Answers

A:

Public cloud services are offered by a provider to multiple customers on shared physical infrastructure, with logical isolation of customer resources.

B:

Multicloud refers to use of multiple cloud providers; hybrid describes integration between public cloud and a private/on-premises environment.

D:

Private clouds serve one organization and emphasize control over infrastructure and operating boundaries, whether hosted on premises or by a provider.

Question 197

A provider states it offers an uptime target and the procurement team wants the formal commitment.

Which observed result would best demonstrate that the intended requirement has been met?

  1. Load and node-failure tests separately show throughput growth and continuity during a lost instance.
  2. The architect documents the SLA and records an availability target with measurement rules appropriate to the service.
  3. A failure test separates host failover from a region-unavailability drill with explicit recovery metrics.
  4. A regional-failure drill measures restoration within the agreed RTO and data loss within the agreed RPO.

Explanation

Correct Answer

B:

An SLA defines the provider's committed service level and related terms; application uptime also depends on customer architecture.

Incorrect Answers

A:

That observation could confirm that load and node-failure tests separately show throughput growth and continuity during a lost instance, but it does not establish that the architect documents the SLA and records an availability target with measurement rules appropriate to the service.

C:

That observation could confirm that a failure test separates host failover from a region-unavailability drill with explicit recovery metrics, but it does not establish that the architect documents the SLA and records an availability target with measurement rules appropriate to the service.

D:

That observation could confirm that a regional-failure drill measures restoration within the agreed RTO and data loss within the agreed RPO, but it does not establish that the architect documents the SLA and records an availability target with measurement rules appropriate to the service.

Question 198

A company installs its own custom database engine on a rented VM and says the cloud provider now owns database patching.

Which cloud or Azure approach best meets the stated requirement?

  1. Treat database engine administration on a customer-managed VM as the customer's IaaS responsibility.
  2. Use PaaS with provider-managed platform maintenance and customer-managed code and data.
  3. Leave business-data authorization and database user access rules under customer control.
  4. Retain customer governance for user identity, access grants and business data in SaaS.

Explanation

Correct Answer

A:

Installing software on IaaS does not turn it into PaaS; the customer maintains software inside the guest. This is the implementation most closely tied to the actual requirement.

Incorrect Answers

B:

PaaS transfers the operating system/runtime maintenance burden to the provider while preserving customer responsibility for application logic and identity.

C:

A managed database provider operates underlying service infrastructure while the customer controls schemas, users, and access to its records.

D:

SaaS reduces customer platform administration but does not automatically decide the organization's authorized users.

Question 199

A merchant wants customer requests distributed among several healthy backend servers.

Which observed result would best demonstrate that the intended requirement has been met?

  1. A simulated zone outage leaves the application reachable through healthy instances in other zones.
  2. Planned maintenance of a single hosting node leaves healthy replicas serving users.
  3. A load test demonstrates extra instances join the service and per-instance pressure falls.
  4. Stopping one backend VM does not prevent test clients from reaching the service through other healthy endpoints.

Explanation

Correct Answer

D:

Load balancing distributes traffic to eligible healthy endpoints; a standalone server remains a single point of failure.

Incorrect Answers

A:

That observation could confirm that a simulated zone outage leaves the application reachable through healthy instances in other zones, but it does not establish that stopping one backend VM does not prevent test clients from reaching the service through other healthy endpoints.

B:

That observation could confirm that planned maintenance of a single hosting node leaves healthy replicas serving users, but it does not establish that stopping one backend VM does not prevent test clients from reaching the service through other healthy endpoints.

C:

That observation could confirm that a load test demonstrates extra instances join the service and per-instance pressure falls, but it does not establish that stopping one backend VM does not prevent test clients from reaching the service through other healthy endpoints.

Question 200

A company wants a ready-to-use hosted productivity application with updates managed by the vendor.

Which cloud or Azure approach best meets the stated requirement?

  1. Choose PaaS as the balance of custom application ownership and provider-managed platform.
  2. Treat the workload as IaaS with customer-managed software inside the guest.
  3. Choose SaaS for complete provider-hosted software consumed by end users.
  4. Choose a service model solely by sticker price without checking who manages software, runtime or guest operating system.

Explanation

Correct Answer

C:

SaaS delivers finished applications; the customer configures tenant access and data use rather than building the software stack. This is the implementation most closely tied to the actual requirement.

Incorrect Answers

A:

PaaS allows coding and deployment while delegating much of server operation to the provider.

B:

A self-installed database or web server does not become provider-managed PaaS merely because it runs in cloud infrastructure.

D:

Continuing the deficient manual process leaves the stated need unresolved. SaaS delivers finished applications; the customer configures tenant access and data use rather than building the software stack.

Microsoft Azure Fundamentals (AZ-900) Practice Tests

Purchase Individually

  • Premium File

    482 Questions & Answers
    Last Update: Nov 4, 2026

    $76.99
    $69.99
  • Training Course

    85 Lectures

    $43.99
    $39.99
  • Study Guide

    425 Pages

    $43.99
    $39.99

SPECIAL OFFER: GET 10% OFF
This is ONE TIME OFFER

You save
10%

Enter Your Email Address to Receive Your 10% Off Discount Code

SPECIAL OFFER: GET 10% OFF

You save
10%

Use Discount Code:

A confirmation link was sent to your e-mail.

Please check your mailbox for a message from support@examlabs.com and follow the directions.

Download Free Demo of VCE Exam Simulator

Experience Avanset VCE Exam Simulator for yourself.

Simply submit your email address below to get started with our interactive software demo of your free trial.

  • Realistic exam simulation and exam editor with preview functions
  • Whole exam in a single file with several different question types
  • Customizable exam-taking mode & detailed score reports