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
You save $69.98
Architecture rationale review—PostgreSQL read replica: Add supported PostgreSQL Flexible Server read replicas with suitable replication-lag expectations. Which explanation accurately distinguishes this capability?
Read replicas offload eligible read-only queries but are asynchronously updated from the writer. It is material here because an application uses PostgreSQL for writes but analysts overload the writer with complex read-only reports.
Partition elimination and aligned indexes can reduce scans when predicates match partition keys. That statement describes SQL database partitioning, while the question asks how PostgreSQL read replica works: Read replicas offload eligible read-only queries but are asynchronously updated from the writer.
Ledger records cryptographically verifiable tamper-evidence for supported relational tables. Read replicas offload eligible read-only queries but are asynchronously updated from the writer.
Flexible Server offers managed open-source PostgreSQL with supported extensions and deployment options. It belongs to PostgreSQL flexible server. The stated architecture instead relies on: Read replicas offload eligible read-only queries but are asynchronously updated from the writer.
Architecture review: GPU VM and Memory optimized VM. Requirement A: A model inference service needs supported GPU acceleration and low response latency. Requirement B: An in-memory analytics engine repeatedly pages memory on a standard VM size. Which complete two-part proposal satisfies both constraints?
For requirement A, GPU instances provide accelerators not present on standard CPU-only VMs. For requirement B, memory-optimized sizes offer higher RAM-to-vCPU ratio for large in-memory working sets. Verify both independently: The model uses available GPU memory and meets measured inference latency; Paging falls and the working data fits within the selected RAM capacity.
Both proposed halves address other use cases: Azure Functions durable workflow and General purpose VM. The requested pair is GPU VM and Memory optimized VM.
One action is correct. The other targets Azure Functions durable workflow, leaving GPU VM unsolved. Its necessary action is: Use an Azure GPU-capable VM family compatible with the model runtime.
The mismatched half is General purpose VM; Memory optimized VM instead needs: Select a memory-optimized Azure VM family after measuring peak resident memory.
Architecture review: Managed-identity scope and Vault RBAC least privilege. Requirement A: A workload requires write access to one blob container but must not inherit owner rights over the subscription. Requirement B: A team member needs read-only access to selected vault secrets but should not grant access policies. Which complete two-part proposal satisfies both constraints?
For requirement A, data-plane roles control access to blob content and least-scope assignment reduces unnecessary privileges. For requirement B, Key Vault RBAC data permissions differ from vault resource-management permissions. Verify both independently: Container writes succeed while administrative ownership actions remain denied; The user retrieves authorized secret values but cannot alter vault access configuration.
Both proposed halves address other use cases: Conditional MFA and App registration API scopes. The requested pair is Managed-identity scope and Vault RBAC least privilege.
One action is correct. The other targets Conditional MFA, leaving Managed-identity scope unsolved. Its necessary action is: Grant a suitable storage data role at the narrow supported resource scope.
Data-plane roles control access to blob content and least-scope assignment reduces unnecessary privileges. The other proposed action (define exposed API permissions in the app registration and require consent as appropriate) addresses a different operation. Key Vault RBAC data permissions differ from vault resource-management permissions.
For Private inbound App Service, operations are closing a change that addressed: An App Service must accept requests only through the corporate private network. Which post-change signal belongs in the acceptance record?
Private endpoints provide a private inbound path for supported App Service access. Confirm using the measurable result: Corporate users connect using private DNS and public clients are rejected.
Private endpoints attach supported PaaS access to a private IP and require DNS resolution. The Private endpoint test checks the data source resolves to a VNet private address and public access is denied; the relevant outcome is Corporate users connect using private DNS and public clients are rejected.
Edge WAF can inspect and filter HTTP requests before the proxy forwards them to origins. The offered observation would demonstrate Front Door WAF edge: A permitted request reaches its origin while a malicious test pattern is blocked at the edge. The current validation is different: Corporate users connect using private DNS and public clients are rejected.
ExpressRoute links customer WAN to Azure backbone privately through provider connectivity. Those observations validate ExpressRoute private circuit. They do not establish the required Private inbound App Service result: Corporate users connect using private DNS and public clients are rejected.
A global enterprise requires central baseline policies applied to every subscription but autonomous project teams manage resources. Which Azure architecture would most directly achieve the required result: A new subscription inherits the applicable central compliance policy?
Management groups supply governance scope above subscriptions without moving resources into a single subscription. The selected architecture is build a management-group hierarchy with appropriate policy and RBAC inheritance; confirm the result by checking that a new subscription inherits the applicable central compliance policy.
Landing zones combine platform foundations with repeatable application subscription governance. Its design target is Landing zones. The stated requirement instead concerns: A global enterprise requires central baseline policies applied to every subscription but autonomous project teams manage resources.
RBAC assignments inherit downward so narrower scope reduces unintended changes. The current need is: A global enterprise requires central baseline policies applied to every subscription but autonomous project teams manage resources.
Cost exports provide detailed recurring usage data for downstream allocation reporting. It does not solve the specified Management group hierarchy condition.
A remote facility must move 500 TB to Azure but WAN bandwidth is too slow for the deadline. An engineer must deliver the following acceptance evidence: The imported objects reconcile against checksums and manifests. What design choice is defensible?
Physical data transfer can shorten initial bulk migration when network transfer is impractical. The selected architecture is use an appropriate Azure Data Box physical transfer service; confirm the result by checking that the imported objects reconcile against checksums and manifests.
Offline migration can simplify cutover when application downtime is acceptable. It does not solve the specified Storage bulk offline transfer condition.
PostgreSQL migration requires extension, version and connectivity compatibility checks. The current need is: A remote facility must move 500 TB to Azure but WAN bandwidth is too slow for the deadline.
Hidden dependencies can break after migration even when VM replication succeeds. Its design target is Application dependency mapping. The stated requirement instead concerns: A remote facility must move 500 TB to Azure but WAN bandwidth is too slow for the deadline.
Architecture review: SQL serverless idle cost and SQL Hyperscale growth. Requirement A: A dev/test SQL database is idle for long periods but bursts during a short daily test window. Requirement B: One operational SQL database must scale beyond the capacity limits of ordinary General Purpose storage tiers. Which complete two-part proposal satisfies both constraints?
For requirement A, serverless compute can scale usage and pause eligible databases to reduce idle cost. For requirement B, hyperscale decouples storage scale from compute and supports large database sizes. Verify both independently: Compute charges drop during idle periods and the test workload resumes within tolerated delay; A growth test reaches the required storage target with the expected compute and backup architecture.
The option covers one requirement but substitutes SQL read scaling for SQL Hyperscale growth. The missing acceptance result is: A growth test reaches the required storage target with the expected compute and backup architecture.
Hyperscale decouples storage scale from compute and supports large database sizes. The other proposed action (plan supported vertical compute scaling and evaluate tier limits and transition impacts) does not implement the first control. Serverless compute can scale usage and pause eligible databases to reduce idle cost.
This combines SQL compute scaling and SQL read scaling, which do not satisfy the distinct requirements for SQL serverless idle cost and SQL Hyperscale growth.
Architecture review: Cosmos multi-write availability and Traffic Manager DNS routing. Requirement A: A global application must continue accepting writes from multiple regions when one region fails. Requirement B: A TCP application requires global endpoint selection using DNS without a global L7 reverse proxy. Which complete two-part proposal satisfies both constraints?
For requirement A, multi-region writes allow eligible regions to accept writes but require handling conflicts. For requirement B, traffic Manager is DNS-based and clients connect directly to the returned service endpoint. Verify both independently: After one region becomes unavailable another authorized region continues to accept writes; DNS responses prefer the healthy endpoint as monitoring marks another unavailable.
Multi-region writes allow eligible regions to accept writes but require handling conflicts. Traffic Manager is DNS-based and clients connect directly to the returned service endpoint. The proposed steps (enable AKS cluster autoscaler within approved node pool bounds; choose zone-redundant storage where the account and service support it) address other capabilities instead of this required pair.
Traffic Manager is DNS-based and clients connect directly to the returned service endpoint. The other proposed action (enable AKS cluster autoscaler within approved node pool bounds) does not implement the first control. Multi-region writes allow eligible regions to accept writes but require handling conflicts.
Multi-region writes allow eligible regions to accept writes but require handling conflicts. The other proposed action (choose zone-redundant storage where the account and service support it) addresses a different operation. Traffic Manager is DNS-based and clients connect directly to the returned service endpoint.
A solution must address both Notification routing and CPU threshold alerts. First requirement: Critical infrastructure alerts need pager escalation while routine capacity warnings go only to email. Second requirement: A fleet needs immediate warning when sustained processor utilization exceeds its target without reading logs manually. Select TWO recommendations that meet these independent needs.
For the CPU threshold alerts half of this design, metrics are low-latency numeric series and alert evaluation can trigger actions independently of a log query. Check: A simulated sustained high-CPU period fires the metric alert and notifies the on-call team. The other half concerns Notification routing.
For the Notification routing half of this two-part design, action groups define notifications and automated responses and can be reused by alert rules. Verify: A simulated critical alert pages the incident team while a capacity warning sends email only. The other half concerns CPU threshold alerts.
AMPLS provides private connectivity to supported Azure Monitor ingestion and query endpoints. Its own validation would be: Monitoring ingestion succeeds using private DNS resolution while public network access is restricted. The current paired requirement is Notification routing and CPU threshold alerts.
Resource Health reports the platform-side availability history of individual Azure resources. That plan is appropriate for VM resource health. The two controls sought here are Notification routing and CPU threshold alerts.
Storage-based diagnostic archives and immutability serve compliance preservation rather than interactive KQL search. Its own validation would be: A retained log blob rejects unauthorized deletion during its compliance retention window. The current paired requirement is Notification routing and CPU threshold alerts.
During technical sign-off for Private observability, the architect selects: Use Azure Monitor Private Link Scope with supported private endpoints and DNS integration. Which underlying behavior should be documented?
AMPLS provides private connectivity to supported Azure Monitor ingestion and query endpoints. It is material here because a regulated environment must ingest monitoring data from isolated private networks without exposing public workspace ingestion endpoints.
Analytics and longer-term retention options have different query and storage cost characteristics. AMPLS provides private connectivity to supported Azure Monitor ingestion and query endpoints.
Container insights exposes Kubernetes workload health and container logs beyond generic virtual-machine statistics. It belongs to AKS container diagnosis. The stated architecture instead relies on: AMPLS provides private connectivity to supported Azure Monitor ingestion and query endpoints.
Entra diagnostic export makes identity sign-in events available for central retention and KQL correlation. That statement describes Identity-signin monitoring, while the question asks how Private observability works: AMPLS provides private connectivity to supported Azure Monitor ingestion and query endpoints.
Architecture review: Quotas and limits and Role assignment scope. Requirement A: A proposed subscription hosts many public IPs and compute cores beyond regional quota. Requirement B: An engineer only operates one network resource group but has Contributor for the whole subscription. Which complete two-part proposal satisfies both constraints?
For requirement A, Azure service quotas are often scoped by subscription and region and can constrain deployability. For requirement B, RBAC assignments inherit downward so narrower scope reduces unintended changes. Verify both independently: A capacity forecast and quota decision support the planned deployment without surprise allocation errors; The engineer can change the assigned network group but not another group's resources.
Both proposed halves address other use cases: Defender compliance and Standardized IaC. The requested pair is Quotas and limits and Role assignment scope.
RBAC assignments inherit downward so narrower scope reduces unintended changes. The other proposed action (use Microsoft Defender for Cloud regulatory compliance and security posture assessments) does not implement the first control. Azure service quotas are often scoped by subscription and region and can constrain deployability.
Azure service quotas are often scoped by subscription and region and can constrain deployability. The other proposed action (use Azure Bicep modules and deployment pipelines with approval gates) addresses a different operation. RBAC assignments inherit downward so narrower scope reduces unintended changes.
Please check your mailbox for a message from support@examlabs.com and follow the directions.