Pass Microsoft AZ-305 Exam in First Attempt Easily
Real Microsoft AZ-305 Exam Questions, Accurate & Verified Answers As Experienced in the Actual Test!

Verified by experts
3 products

You save $69.98

AZ-305 Premium Bundle

  • Premium File 382 Questions & Answers
  • Last Update: Nov 20, 2026
  • Training Course 87 Lectures
  • Study Guide 933 Pages
$79.99 $149.97

Purchase Individually

  • Premium File

    382 Questions & Answers
    Last Update: Nov 20, 2026

    $76.99
    $69.99
  • Training Course

    87 Lectures

    $43.99
    $39.99
  • Study Guide

    933 Pages

    $43.99
    $39.99

Designing Microsoft Azure Infrastructure Solutions AZ-305 Practice Test 10

Question 90

The acceptance criterion for Load Balancer internal L4 comes from this issue: A TCP database-proxy cluster needs one private IP balancing connections to multiple healthy nodes. Which measured outcome is the strongest proof?

  1. The test identifies the blocking NSG rule or unreachable next hop before changing production routes
  2. The approved subnet accesses storage and unauthorized internet clients are blocked
  3. Private IP traffic flows between the peered VNets with appropriate NSG permissions
  4. An unhealthy proxy node stops receiving new connections through the internal VIP

Explanation

Correct Answer

D:

Azure Load Balancer routes Layer 4 traffic inside a VNet without parsing HTTP URLs. Confirm using the measurable result: An unhealthy proxy node stops receiving new connections through the internal VIP.

Incorrect Answers

A:

Network Watcher diagnostic tools identify effective network behavior and filtering decisions. The Network Watcher diagnosis test checks the test identifies the blocking NSG rule or unreachable next hop before changing production routes; the relevant outcome is An unhealthy proxy node stops receiving new connections through the internal VIP.

B:

Service endpoints keep the platform public endpoint while allowing VNet-based service firewall authorization. The offered observation would demonstrate Service endpoint: The approved subnet accesses storage and unauthorized internet clients are blocked. The current validation is different: An unhealthy proxy node stops receiving new connections through the internal VIP.

C:

Peering connects VNet address spaces through the Microsoft backbone without creating transitive routing by default. Those observations validate VNet peering. They do not establish the required Load Balancer internal L4 result: An unhealthy proxy node stops receiving new connections through the internal VIP.

Question 91

The SQL database PITR review proposes: Use managed Azure SQL point-in-time restore to a new database within backup retention. Which Azure behavior is the technical justification for this decision?

  1. Geo-restore is backup-based recovery with different RTO/RPO than active geo-replication
  2. Azure SQL automatic full differential and log backups support restoring database state at a selected time
  3. Failback is a separate operational sequence after workloads run in the recovery site
  4. Share snapshots preserve file-share point-in-time state for supported restoration

Explanation

Correct Answer

B:

Azure SQL automatic full differential and log backups support restoring database state at a selected time. It is material here because a deployment accidentally truncates a transaction table and the database must be recovered before the change.

Incorrect Answers

A:

Geo-restore is backup-based recovery with different RTO/RPO than active geo-replication. Azure SQL automatic full differential and log backups support restoring database state at a selected time.

C:

Failback is a separate operational sequence after workloads run in the recovery site. It belongs to Failback after recovery. The stated architecture instead relies on: Azure SQL automatic full differential and log backups support restoring database state at a selected time.

D:

Share snapshots preserve file-share point-in-time state for supported restoration. That statement describes Azure Files snapshots, while the question asks how SQL database PITR works: Azure SQL automatic full differential and log backups support restoring database state at a selected time.

Question 92

The design team completed work on ADF incremental copy. Its governing requirement was: Only rows changed since the last successful extraction should be copied from a large source database. Identify the conclusive verification result.

  1. A second run copies only new or modified rows and advances the checkpoint after success
  2. The scheduled copy reads on-premises data and lands it in the lake without inbound firewall exceptions
  3. Authorized Fabric consumers query the shared curated data with governed access
  4. The pipeline reads the database through the authorized private path while public access fails

Explanation

Correct Answer

A:

A stored watermark limits reads to changes after the last successful processed interval. Confirm using the measurable result: A second run copies only new or modified rows and advances the checkpoint after success.

Incorrect Answers

B:

Self-hosted integration runtime initiates supported outbound communications to access private sources. The distinguishing check is: A second run copies only new or modified rows and advances the checkpoint after success.

C:

OneLake offers a common analytics data foundation for supported Fabric workloads. The Fabric OneLake analytics test checks authorized Fabric consumers query the shared curated data with governed access; the relevant outcome is A second run copies only new or modified rows and advances the checkpoint after success.

D:

Hybrid/private integration requires a reachable private network path rather than open public DB exposure. The offered observation would demonstrate Private integration access: The pipeline reads the database through the authorized private path while public access fails. The current validation is different: A second run copies only new or modified rows and advances the checkpoint after success.

Question 93

Operations wants reproducible Azure environment deployments from version-controlled definitions. An engineer must deliver the following acceptance evidence: The same versioned module deploys matching compliant resources across test environments. What design choice is defensible?

  1. Place regulated workloads in designated subscriptions under suitable management groups
  2. Build an Azure Workbook over centrally collected compliance and resource data
  3. Create a policy remediation task with suitable identity permissions for supported deployIfNotExists rules
  4. Use Azure Bicep modules and deployment pipelines with approval gates

Explanation

Correct Answer

D:

Infrastructure as code declares desired resources and allows review and repeat deployment. The selected architecture is use Azure Bicep modules and deployment pipelines with approval gates; confirm the result by checking that the same versioned module deploys matching compliant resources across test environments.

Incorrect Answers

A:

Subscriptions are billing and resource governance boundaries with independent quotas and RBAC assignment scopes. That is useful for Subscription isolation (A regulated subsidiary needs separate billing and policy boundaries from the parent business unit). It does not meet the current Standardized IaC requirement: Operations wants reproducible Azure environment deployments from version-controlled definitions.

B:

Workbooks provide interactive analytical reports rather than the policy enforcement mechanism itself. It does not solve the specified Standardized IaC condition.

C:

DeployIfNotExists requires remediation for already deployed resources where appropriate. The relevant requirement is: Operations wants reproducible Azure environment deployments from version-controlled definitions.

Question 94

For Event Hubs telemetry, operations are closing a change that addressed: Thousands of kiosks upload constant high-volume telemetry that several analytic consumers replay. Which post-change signal belongs in the acceptance record?

  1. The user session persists after a web worker is intentionally removed
  2. Restarting a worker does not restart already completed business steps
  3. Independent consumer groups process events at different rates using their own offsets
  4. Two deliveries of the same order ID create only one invoice

Explanation

Correct Answer

C:

Event Hubs provides a replayable ordered stream within partitions for large event volumes. Confirm using the measurable result: Independent consumer groups process events at different rates using their own offsets.

Incorrect Answers

A:

Shared session state survives an individual stateless app worker replacement. The Azure Cache for Redis session test checks the user session persists after a web worker is intentionally removed; the relevant outcome is Independent consumer groups process events at different rates using their own offsets.

B:

Durable orchestrations checkpoint progress and coordinate stateful serverless workflows. The offered observation would demonstrate Durable Functions orchestration: Restarting a worker does not restart already completed business steps. The current validation is different: Independent consumer groups process events at different rates using their own offsets.

D:

At-least-once message delivery requires application logic to deduplicate business side effects. The distinguishing check is: Independent consumer groups process events at different rates using their own offsets.

Question 95

A deployment accidentally truncates a transaction table and the database must be recovered before the change. Identify the most suitable implementation for an outcome where a restored database contains transactions immediately preceding the truncation.

  1. Use Flexible Server point-in-time restore within supported backup-retention limits
  2. Use replicated disaster recovery architecture and separately retain backups for point-in-time recovery
  3. Use managed Azure SQL point-in-time restore to a new database within backup retention
  4. Plan ASR recovery VNet, IP addressing, NSGs, routing and DNS before failover

Explanation

Correct Answer

C:

Azure SQL automatic full differential and log backups support restoring database state at a selected time. The selected architecture is use managed Azure SQL point-in-time restore to a new database within backup retention; confirm the result by checking that a restored database contains transactions immediately preceding the truncation.

Incorrect Answers

A:

Managed PostgreSQL backups support rebuilding a server at a selected valid recovery time. The relevant requirement is: A deployment accidentally truncates a transaction table and the database must be recovered before the change.

B:

Nightly backups alone do not provide continuous workload replication or low recovery time. It does not solve the specified SQL database PITR condition.

D:

Compute replication does not automatically guarantee complete application network reachability. Its design target is Azure-to-Azure ASR networking. The stated requirement instead concerns: A deployment accidentally truncates a transaction table and the database must be recovered before the change.

Question 96

A solution must address both Memory optimized VM and Compute optimized VM. First requirement: An in-memory analytics engine repeatedly pages memory on a standard VM size. Second requirement: A stateless batch worker is CPU-bound with little disk IO or memory pressure. Select TWO recommendations that meet these independent needs.

  1. Choose an appropriately sized general-purpose Azure VM with supporting disks and network
  2. Use Azure Batch pools, tasks and job orchestration
  3. Choose a compute-optimized VM family sized for sustained CPU throughput
  4. Select a memory-optimized Azure VM family after measuring peak resident memory
  5. Use an Azure GPU-capable VM family compatible with the model runtime

Explanation

Correct Answers

C:

For the Compute optimized VM half of this design, compute-optimized instances prioritize vCPU capacity over excess memory. Check: The same batch finishes within the required CPU time without memory bottlenecks. The other half concerns Memory optimized VM.

D:

For the Memory optimized VM half of this two-part design, memory-optimized sizes offer higher RAM-to-vCPU ratio for large in-memory working sets. Verify: Paging falls and the working data fits within the selected RAM capacity. The other half concerns Compute optimized VM.

Incorrect Answers

A:

General-purpose VMs balance CPU and memory while retaining OS-level control. The corresponding task is General purpose VM: A legacy Windows service requires full OS control and custom drivers but has moderate CPU and memory needs. It is not one of the two requested controls (Memory optimized VM and Compute optimized VM).

B:

Azure Batch schedules large-scale parallel workloads across managed compute pools. That plan is appropriate for Azure Batch scheduling. The two controls sought here are Memory optimized VM and Compute optimized VM.

E:

GPU instances provide accelerators not present on standard CPU-only VMs. The corresponding task is GPU VM: A model inference service needs supported GPU acceleration and low response latency. It is not one of the two requested controls (Memory optimized VM and Compute optimized VM).

Question 97

Thousands of kiosks upload constant high-volume telemetry that several analytic consumers replay. A deployment must satisfy this operational criterion: Independent consumer groups process events at different rates using their own offsets. Which plan is appropriate?

  1. Choose Logic Apps workflows with supported connector authentication
  2. Use Azure Front Door rule sets for supported redirects and response-header actions
  3. Configure API Management rate limiting or quota policies at the appropriate scope
  4. Use Event Hubs partitions and consumer groups for streaming ingestion

Explanation

Correct Answer

D:

Event Hubs provides a replayable ordered stream within partitions for large event volumes. The selected architecture is use Event Hubs partitions and consumer groups for streaming ingestion; confirm the result by checking that independent consumer groups process events at different rates using their own offsets.

Incorrect Answers

A:

Logic Apps offers managed integration workflows and prebuilt service connectors. It does not solve the specified Event Hubs telemetry condition.

B:

Front Door rule sets change eligible request and response behavior at the edge according to match conditions. Its design target is Front Door rules engine. The stated requirement instead concerns: Thousands of kiosks upload constant high-volume telemetry that several analytic consumers replay.

C:

APIM can enforce per-client API consumption controls before requests reach backends. The relevant requirement is: Thousands of kiosks upload constant high-volume telemetry that several analytic consumers replay.

Question 98

A solution must address both Web PubSub real time and Azure Cache for Redis session. First requirement: A collaboration app must send real-time server-to-client notifications to many concurrent browsers. Second requirement: A scaled-out web app stores user sessions on individual workers and loses sessions on failover. Select TWO recommendations that meet these independent needs.

  1. Configure API Management rate limiting or quota policies at the appropriate scope
  2. Reference Key Vault secrets using workload identity from the application
  3. Move session state to a shared Azure Redis-compatible cache or durable store
  4. Use Azure Web PubSub with authenticated WebSocket clients and appropriate hub design
  5. Use Service Bus topics with subscriptions when reliable broker delivery is needed

Explanation

Correct Answers

C:

For the Azure Cache for Redis session half of this design, shared session state survives an individual stateless app worker replacement. Check: The user session persists after a web worker is intentionally removed. The other half concerns Web PubSub real time.

D:

For the Web PubSub real time half of this two-part design, web PubSub provides managed WebSocket connection fan-out so application servers need not hold all client sessions. Verify: Two connected clients immediately receive the authorized published event. The other half concerns Azure Cache for Redis session.

Incorrect Answers

A:

APIM can enforce per-client API consumption controls before requests reach backends. Its own validation would be: A burst beyond the quota is throttled while permitted traffic proceeds. The current paired requirement is Web PubSub real time and Azure Cache for Redis session.

B:

Central secret retrieval separates sensitive credentials from source and build artifacts. This alternative solves Key Vault app secrets, while the question separately requires Web PubSub real time and Azure Cache for Redis session.

E:

Topics fan out a message to independently managed subscriptions with durable handling. Its own validation would be: All three subscribers receive an order event without direct calls from checkout. The current paired requirement is Web PubSub real time and Azure Cache for Redis session.

Question 99

The SQL customer-managed keys review proposes: Configure customer-managed transparent data encryption protector where supported. Which Azure behavior is the technical justification for this decision?

  1. Flexible Server offers managed open-source PostgreSQL with supported extensions and deployment options
  2. Ledger records cryptographically verifiable tamper-evidence for supported relational tables
  3. Temporal tables retain row history and support as-of queries independently of whole-database backups
  4. CMK encryption protects database data at rest while keys are managed under customer control

Explanation

Correct Answer

D:

CMK encryption protects database data at rest while keys are managed under customer control. It is material here because compliance requires the SQL database encryption protector to use customer-controlled key lifecycle in Key Vault.

Incorrect Answers

A:

Flexible Server offers managed open-source PostgreSQL with supported extensions and deployment options. CMK encryption protects database data at rest while keys are managed under customer control.

B:

Ledger records cryptographically verifiable tamper-evidence for supported relational tables. CMK encryption protects database data at rest while keys are managed under customer control.

C:

Temporal tables retain row history and support as-of queries independently of whole-database backups. It belongs to SQL temporal tables. The stated architecture instead relies on: CMK encryption protects database data at rest while keys are managed under customer control.

Question 100

A business-critical file repository must survive an availability-zone outage without a regional failover. Which Azure architecture would most directly achieve the required result: Reading and writing a test blob continues while one zone is unavailable?

  1. Add appropriate Cosmos DB read regions and choose the fitting consistency level
  2. Use a supported zone-redundant Application Gateway WAF tier
  3. Use zone-aware node pools and workload spread/anti-affinity policies
  4. Choose zone-redundant storage where the account and service support it

Explanation

Correct Answer

D:

ZRS synchronously replicates data across primary-region zones. The selected architecture is choose zone-redundant storage where the account and service support it; confirm the result by checking that reading and writing a test blob continues while one zone is unavailable.

Incorrect Answers

A:

Read regions bring replicas closer to users without inherently adding multi-write semantics. The current need is: A business-critical file repository must survive an availability-zone outage without a regional failover.

B:

Application Gateway offers L7 proxy capabilities with WAF unlike a basic L4 load balancer. That is useful for Application Gateway zone redundancy (A web workload requires HTTP routing and WAF across availability zones). It does not meet the current Blob ZRS availability requirement: A business-critical file repository must survive an availability-zone outage without a regional failover.

C:

Zone-distributed nodes plus suitable Kubernetes scheduling reduce single-zone concentration. It does not solve the specified Blob ZRS availability condition.

Designing Microsoft Azure Infrastructure Solutions (AZ-305) Practice Tests

Purchase Individually

  • Premium File

    382 Questions & Answers
    Last Update: Nov 20, 2026

    $76.99
    $69.99
  • Training Course

    87 Lectures

    $43.99
    $39.99
  • Study Guide

    933 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