McAfee-Secured Website

Exam Bundle

Exam Code: AZ-104

Exam Name Microsoft Azure Administrator

Certification Provider: Microsoft

Corresponding Certification: Microsoft Certified: Azure Administrator Associate

Microsoft AZ-104 Bundle $44.99

100% Microsoft AZ-104 Guaranteed Success With Testking Training Materials

Get Real AZ-104 Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

  • Questions & Answers

    AZ-104 Questions & Answers

    606 Questions & Answers

    The ultimate exam preparation tool, Questions & Answers mirrors the actual AZ-104 exam while allowing you to hone and measure every aspect of your test-taking ability.

  • AZ-104 Video Course

    AZ-104 Video Course

    132 Video Lectures

    AZ-104 Video Course is developed by Microsoft Professionals to help you pass the AZ-104 exam.

    Description

    This course will improve your knowledge and skills required to pass Microsoft Azure Administrator exam.
  • Study Guide

    AZ-104 Study Guide

    458 PDF Pages

    Developed by industry experts, this 458-page guide spells out in painstaking detail all of the information you need to ace AZ-104 exam.

Microsoft Azure Administrator (AZ-104) Practice Tests

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Total Cost: $194.97
Bundle Price: $149.98

Purchase Individually

  • Questions & Answers

    Questions & Answers

    606 Questions

    $124.99
  • AZ-104 Video Course

    Video Course

    132 Video Lectures

    $39.99
  • Study Guide

    Study Guide

    458 PDF Pages

    $29.99

Microsoft AZ-104 Practice Test 27

Question 260

A Bicep template receives an administrator credential as a secure parameter. A developer then declares a normal output containing the plaintext password. Which risk should the reviewer flag?

  1. The secure parameter automatically encrypts every ordinary output referencing it.
  2. Secrets must not be exposed through ordinary deployment outputs; use supported secure handling and avoid emitting credentials.
  3. Store the plaintext password in a normal output named securePassword, and the output identifier itself prevents deployment-history disclosure.
  4. Mark the input secure but leave the output a normal string, and the input decorator protects its downstream output value.

Explanation

Correct Answer

B:

Deployment outputs can be visible through deployment records and pipelines. Secure input alone does not make a plaintext output safe.

Incorrect Answers

A:

A secure parameter protects the input handling but does not justify echoing the value in output.

C:

An output name has no secret-protection semantics. A normal deployment output can expose its value through deployment operations or pipeline logs.

D:

The input’s secure designation does not make an ordinary plaintext output safe to disclose. Avoid exporting the credential.

Question 261

A team updates the scale-set model but must validate the new image on selected instances before updating production workers. Which upgrade strategy gives operators explicit control?

  1. Set the upgrade policy to Automatic and use a calendar note to decide when VMSS instances may apply the changed model.
  2. Use automatic upgrades and operators can always hold every instance unchanged indefinitely.
  3. Use a supported manual VMSS upgrade policy and roll updates to instances under the controlled deployment plan.
  4. Change the scale-set deployment history name to ManualUpgrade without altering the instance upgrade policy.

Explanation

Correct Answer

C:

Manual upgrade mode separates model changes from instance application. The team can stage and validate the rollout explicitly.

Incorrect Answers

A:

Automatic upgrades do not provide the desired per-instance, operator-controlled rollout after model changes.

B:

Automatic upgrade behavior can apply changes without the same manual per-instance control.

D:

Deployment metadata naming does not change how VMSS propagates a model update to instances.

Question 262

A VM Scale Set runs a stateless app that should continue operating after an availability-zone outage. The region supports VMSS across zones. Which placement design improves resilience?

  1. Put all instances into a single zone and enable only a CPU alert.
  2. Select supported multi-zone placement for the scale set and verify backend load balancing across healthy zones.
  3. Configure an availability set for VMs in one zone as an equivalent to distributing scale-set instances across multiple zones.
  4. Place the load balancer frontend in a zone-redundant configuration but keep every backend VMSS instance in one zone.

Explanation

Correct Answer

B:

Multi-zone scale set placement distributes compute across independent zone fault domains. Load balancing and application state must also be resilient.

Incorrect Answers

A:

Monitoring a single-zone deployment does not create alternate running capacity after a zone outage.

C:

Within-region fault/update domains do not substitute for using separate availability zones across the VMSS deployment.

D:

Frontend resilience alone does not preserve application compute when all backends share the failed zone.

Question 263

An application uses VMs in an availability set. The team asks whether update domains and fault domains protect against exactly the same event. Which distinction is correct?

  1. Treat fault domains as the rolling planned-update boundary and update domains as physical rack isolation.
  2. Both settings are equivalent labels for the same infrastructure fault grouping.
  3. Fault domains always guarantee replication across different Azure regions.
  4. Fault domains separate infrastructure failure exposure while update domains help stagger planned platform maintenance.

Explanation

Correct Answer

D:

Availability sets use both fault and update domain placement for different resilience concerns. Neither by itself is the same as cross-region DR.

Incorrect Answers

A:

This reverses their roles: fault domains address underlying failure separation, while update domains help stagger planned platform maintenance.

B:

They protect against distinct operational events and should be considered separately for workload availability.

C:

Fault domains within availability sets do not automatically span geographic regions.

Question 264

An administrator increases the size of an Azure VM data disk in the portal but the guest operating system still reports the old filesystem capacity. What else is necessary?

  1. Restart the VM and a guest filesystem automatically expands without a partition or volume change.
  2. Extend the guest partition and filesystem using the supported operating system procedure after the Azure disk change.
  3. Convert the managed disk from Standard to Premium without adjusting its guest filesystem boundary.
  4. Create a managed disk snapshot and attach it as a second disk, and the original partition will grow automatically.

Explanation

Correct Answer

B:

Azure disk capacity and guest partition/file system size are different layers. The guest must recognize and expand the relevant volume safely.

Incorrect Answers

A:

Guest filesystems often need explicit partition and filesystem extension after the Azure disk allocation is enlarged.

C:

Changing disk performance tier does not extend the guest partition or mounted filesystem to consume new capacity.

D:

A snapshot copy is separate storage and does not change the partition size on the existing mounted disk.

Question 265

A rolling scale-set upgrade should stop progressing if newly updated instances fail application health checks, rather than merely seeing that the VM is powered on. What configuration should be examined?

  1. Use only VM power-state health, even when the application can return errors on an otherwise running instance.
  2. Set the rolling upgrade batch size without configuring any application-level health signals.
  3. Set the VMSS model upgrade timeout to zero to skip all health checks.
  4. Configure the supported application health extension or probe and rolling upgrade health policy for the VMSS.

Explanation

Correct Answer

D:

Rolling upgrades can use application-aware health data to gate progress and protect service availability. The probe must reflect the workload's real readiness.

Incorrect Answers

A:

A powered-on VM can have an unhealthy application; application health extension or probe is needed to gate upgrades on real readiness.

B:

Batch limits constrain rollout pace but cannot reliably detect an application that fails after an apparently successful VM update.

C:

Skipping health checks increases failure risk and defeats the requested rollout protection.

Question 266

An event-driven Azure Container Apps job has no running instances between queue messages. It must start an execution when new messages arrive. Which configuration makes the job react to queue work?

  1. Configure an identity with Azure Reader rights on the queue resource but do not create an event trigger or polling rule.
  2. A supported event scaler or job trigger with credentials and scale parameters to detect the queue workload.
  3. Set the Container Apps job minimum execution count to zero without configuring any event trigger or queue scale rule.
  4. Set a Container Apps cron schedule for a fixed time and it reacts immediately to arbitrary new queue messages.

Explanation

Correct Answer

B:

Container Apps scaling depends on configured triggers or metrics, not simply the existence of a container image. The queue scaler must be able to authenticate and read backlog.

Incorrect Answers

A:

Read access alone does not instruct the jobs scaler to watch queue depth and schedule executions.

C:

Zero minimum executions permits idle periods; it does not instruct the runtime to detect queue arrivals. A supported event trigger is still required.

D:

A time-based trigger does not replace an event scaler watching queue messages for on-demand execution.

Question 267

Several microservices should communicate privately inside a common managed Container Apps environment, with consistent workload-level logging and network configuration. Which resource boundary should the designer plan?

  1. Deploy apps into a suitably configured Container Apps environment with the required private networking and observability settings.
  2. Publish all microservice images under the same private Azure Container Registry repository and use repository membership as their internal application network.
  3. Deploy each microservice into a separate Container Apps environment in the same subscription and use subscription scope as their private network boundary.
  4. Associate all microservices with one private DNS zone and use the DNS zone settings as the shared Container Apps logging and ingress configuration.

Explanation

Correct Answer

A:

The environment defines common hosting infrastructure and supported network settings for Container Apps. Image storage and identity are separate concerns.

Incorrect Answers

B:

Repository co-location groups OCI artifacts; an image registry does not provide the shared runtime ingress, virtual network boundary or observability of a Container Apps environment.

C:

Subscription ownership is a management and billing scope, not a single Container Apps managed environment. Separate environments have distinct network/observability boundaries.

D:

Private DNS can resolve eligible service names, but it does not supply managed app revisions, environment-level logging configuration or Container Apps private networking.

Question 268

A deployment pipeline rotated its service principal client secret and several Container Apps can no longer pull images. The platform team wants to avoid storing and rotating per-app registry passwords. What identity design should it evaluate?

  1. Use a subscription-wide Azure Owner role for every app and continue storing a static image-pull password.
  2. Embed the new registry password in the container image layers.
  3. Configure a supported managed identity for image pull and grant narrowly scoped ACR pull permissions.
  4. Make the registry anonymously readable for all images rather than managing workload identities and permissions.

Explanation

Correct Answer

C:

Managed identity avoids static application-managed client secrets for supported registry authentication. Pull role scope and runtime settings still matter.

Incorrect Answers

A:

Owner is overly privileged and does not deliver the credential-free per-workload registry access design requested.

B:

Embedding credentials in images exposes them to image readers and complicates future rotation.

D:

Anonymous access would expose proprietary images and evade the required per-app least-privilege authorization.

Question 269

An Azure Container Instance deployment remains Pending because the requested CPU and memory combination is unsupported in the selected region. The image itself is valid. What should operators check?

  1. Increase only the ACR repository quota while retaining the unsupported ACI CPU and memory requests.
  2. Change the container image tag while retaining the unsupported CPU and memory specification.
  3. Change the container restart policy to Always and the scheduler will provision an unsupported resource profile.
  4. Review supported ACI resource profiles and regional capacity, then adjust CPU/memory requests or placement.

Explanation

Correct Answer

D:

ACI container groups require supported resource requests and placement capacity. A valid image cannot override platform resource availability.

Incorrect Answers

A:

Registry quota concerns image storage and access, not the region’s ACI container group resource profile eligibility.

B:

Image naming does not change ACI regional capacity or whether a requested compute/memory combination is available.

C:

Restart policy affects container process behavior, not whether the initial resource request fits supported regional ACI capacity.

Question 270

A containerized API can scale out rapidly, but the shared database can handle only twenty concurrent app replicas. What platform setting can bound replica growth?

  1. Increase the CPU event-scaler threshold but set no upper replica bound.
  2. Set minimum replicas to twenty while leaving maximum replicas unbounded.
  3. Reduce per-replica HTTP concurrency without configuring the maximum replica count.
  4. Set an appropriate maximum replica count in the Container App scale configuration and monitor database saturation.

Explanation

Correct Answer

D:

maxReplicas constrains the autoscaling upper bound where supported. It helps protect downstream dependencies from unbounded replica growth.

Incorrect Answers

A:

A trigger threshold changes when scaling occurs, not the maximum number of replicas that may be provisioned.

B:

Minimum replicas ensures a baseline but does not cap growth above twenty when demand spikes.

C:

Concurrency changes workload distribution and scaling pressure; it is not an absolute cap on the number of replicas.