AZ-305
Free AZ-305 practice questions
10 questions from our AZ-305 practice tests, each with the correct answer and a written explanation. Read them with the answers shown, or switch to quiz mode and answer first.
Question 1 of 10, Design identity, governance, and monitoring solutions
An application writes about 400 GB per day of verbose trace data to a custom Log Analytics table. Engineers query it a few times a month with simple KQL filters during the first 30 days, after which it must be retained for two years for audit purposes only. Alerting is not required on this data. You need to minimize cost. What should you recommend?
- AKeep the table on the Analytics plan with 30 days of interactive retention and stream a copy of the data to Event Hubs for archiving
- BSwitch the table to the Basic table plan and set its total retention to two years so older data moves to long-term retention Correct answer
- CKeep the table on the Analytics plan, buy a 500 GB/day commitment tier, and set the table's interactive retention to 730 days
- DKeep the table on the Analytics plan, set interactive retention to 90 days, and run a nightly search job that deletes rows older than that
Answer and explanation: The Basic plan charges a much lower ingestion price for data that is queried occasionally and not used for alerts, and total retention can extend to two years in low-cost long-term retention. A commitment tier still bills Analytics rates and two years of interactive retention is the most expensive option. Streaming to Event Hubs adds a pipeline and still pays full Analytics ingestion, and a search job does not delete data or satisfy the two-year requirement.
Question 2 of 10, Design identity, governance, and monitoring solutions
A manufacturer runs Azure workloads in Europe and North America. A regulator requires that platform logs generated by European resources are stored only in Europe and that North American logs stay in North America. A central security team must still be able to run KQL queries across all logs from one place. What should you recommend for the Log Analytics design?
- ADeploy one Log Analytics workspace per geography, send each region's logs to its local workspace, and give the security team cross-workspace queries Correct answer
- BDeploy a single Log Analytics workspace in West Europe and route the diagnostic settings of every resource in both geographies to that workspace
- CDeploy a single Log Analytics workspace in East US and use resource-context RBAC so that European analysts can read only the logs of European resources
- DDeploy one workspace in each geography and configure data export rules that copy every table into a central East US workspace for the security team to query
Answer and explanation: Per-geography workspaces keep each region's data at rest in its own geography, and KQL cross-workspace queries (or a multi-workspace Sentinel view) let the central team query them together without moving data. A single West Europe or East US workspace stores the other geography's logs outside its permitted region; RBAC controls who reads data, not where it is stored. Exporting all tables to a central East US workspace copies European data out of Europe.
Question 3 of 10, Design identity, governance, and monitoring solutions
A company runs 200 Windows and Linux servers in its own datacenters and at another cloud provider. It wants to audit their OS configuration with Azure Policy machine configuration, protect them with Microsoft Defender for Servers, and collect logs with the Azure Monitor Agent, all managed from Azure. What should you recommend first?
- AOnboard the servers to Azure Arc-enabled servers so that they appear as Azure resources that can be governed Correct answer
- BRun an Azure Migrate discovery and assessment so that the servers are imported into the Azure Migrate project
- CInstall the Azure Monitor Agent directly on each server with a workspace ID and key in its configuration file
- DDeploy Azure Local clusters in each datacenter and migrate the servers onto them as Azure Local virtual machines
Answer and explanation: Azure Arc-enabled servers project non-Azure machines into Azure Resource Manager, which is the prerequisite for machine configuration policies, Defender for Servers, and the Azure Monitor Agent on those machines. Azure Local would require re-platforming the servers. The Azure Monitor Agent on non-Azure servers is installed through Arc and does not use workspace keys, and Azure Migrate assesses servers for migration rather than governing them.
Question 4 of 10, Design data storage solutions
A company is moving an on-premises SQL Server 2019 instance to Azure. Its databases use cross-database queries, SQL Server Agent jobs, and Service Broker. The company wants a platform-as-a-service offering with as few code changes as possible. What should you recommend?
- AAzure SQL Database single databases on the General Purpose tier with elastic jobs replacing the Agent jobs
- BAzure SQL Managed Instance on the General Purpose service tier Correct answer
- CSQL Server on an Azure VM with automated patching and backup configured through the SQL IaaS Agent extension
- DAzure SQL Database Hyperscale with read replicas used for the cross-database reporting queries
Answer and explanation: SQL Managed Instance offers near-complete SQL Server compatibility, including cross-database queries, SQL Server Agent, and Service Broker, as a PaaS service. Azure SQL Database single databases, including Hyperscale, do not support cross-database queries or Service Broker in the same way and would require code changes. SQL Server on a VM is compatible but is IaaS, so the company would still manage the OS and SQL Server.
Question 5 of 10, Design data storage solutions
A storage account holds documents that must survive the loss of an availability zone and the loss of the entire primary region. During a regional outage, the application must be able to read the data from the secondary region without waiting for an account failover. Which redundancy option should you recommend?
- ARead-access geo-zone-redundant storage (RA-GZRS) Correct answer
- BRead-access geo-redundant storage (RA-GRS)
- CGeo-zone-redundant storage (GZRS) with customer-managed failover
- DZone-redundant storage (ZRS) combined with object replication to a second account
Answer and explanation: RA-GZRS replicates synchronously across zones in the primary region and asynchronously to a secondary region, and it allows reads from the secondary endpoint at any time. GZRS without read access needs a failover before the secondary can be read. RA-GRS keeps primary copies in a single datacenter, so a zone outage affects it, and ZRS with object replication adds a second account to manage and only copies block blobs.
Question 6 of 10, Design business continuity solutions
A SQL Managed Instance in West Europe hosts 30 databases for a core banking app. The bank needs all databases to fail over together to North Europe during a regional outage with a single, unchanged connection endpoint, and wants the secondary to be usable for read-only reporting. What should you recommend?
- ACreate the Managed Instance link to a SQL Server 2022 instance on an Azure VM in North Europe for read-only access
- BConfigure geo-redundant backup storage and geo-restore each of the 30 databases to a new instance in North Europe
- CConfigure a failover group to a secondary instance in North Europe and use its read-only listener for reporting Correct answer
- DEnable zone redundancy on the Business Critical instance and use its built-in read-only replica for reporting
Answer and explanation: Managed Instance failover groups replicate all user databases to a secondary instance in another region, fail them over together behind stable listener endpoints, and offer a read-only listener for reporting. Geo-restore has high RPO and RTO and must be repeated per database. The Managed Instance link is designed for hybrid SQL Server scenarios, and zone redundancy does not protect against a regional outage.
Question 7 of 10, Design business continuity solutions. Choose 2.
After a ransomware incident at a peer company, a bank wants to ensure that attackers who compromise a backup administrator's account cannot delete Azure VM backups or shorten retention. The protection must not rely on the backup team's own permissions. Which two actions should you recommend? (Choose two.)
- AEnable Cross Region Restore so that the backups can be restored in the secondary region
- BChange the vault's storage redundancy from locally redundant storage to geo-redundant storage
- CApply a CanNotDelete resource lock to the resource group that contains the Recovery Services vault
- DProtect critical vault operations with multi-user authorization using a Resource Guard owned by the security team in a separate subscription Correct answer
- EEnable immutability on the Recovery Services vault and lock it so that it cannot be disabled Correct answer
Answer and explanation: A locked immutable vault prevents operations that delete recovery points or reduce retention, and multi-user authorization requires approval from a Resource Guard controlled by another team before critical operations run. A resource lock can be removed by anyone with Owner rights, and GRS or Cross Region Restore improves availability of backups but does not stop a privileged attacker from deleting them.
Question 8 of 10, Design infrastructure solutions
Apps in a virtual network connect to one Azure SQL Database logical server. Security wants to ensure that compromised workloads cannot copy data to any other Azure SQL server, including servers in attackers' subscriptions, and the database must also be reachable from on-premises over ExpressRoute private peering. What should you recommend?
- AConfigure Azure Firewall with an FQDN rule for *.database.windows.net and route the app subnets through it
- BEnable a Microsoft.Sql service endpoint on the app subnets and add a virtual network rule on the logical server
- CAllow the Sql service tag outbound in the NSGs and set the logical server firewall to allow only the app subnet addresses
- DCreate a private endpoint for the logical server, disable public network access, and block the Sql service tag outbound Correct answer
Answer and explanation: A private endpoint maps one specific server to a private IP that is reachable over ExpressRoute private peering, and blocking the Sql service tag outbound stops connections to any other SQL server. A service endpoint routes traffic to the whole SQL service, so other servers stay reachable, and it does not extend to on-premises. Allowing the Sql service tag or a wildcard FQDN allows connections to every server, including attackers' servers.
Question 9 of 10, Design infrastructure solutions
On-premises servers must resolve the private endpoint names of Azure PaaS services, and Azure VMs must resolve names in the on-premises corp.contoso.com zone. The company wants a managed solution and does not want to run DNS server VMs in Azure. What should you recommend?
- AAzure DNS Private Resolver with an inbound endpoint for on-premises queries and an outbound endpoint with a forwarding ruleset Correct answer
- BAzure private DNS zones linked to the hub network, with on-premises DNS servers forwarding all queries directly to the 168.63.129.16 address
- CAzure public DNS zones that host the private endpoint records with split-horizon views for the on-premises network and Azure VMs
- DA pair of Windows Server DNS VMs in the hub with conditional forwarders to on-premises and to Azure-provided DNS for the private zones
Answer and explanation: Azure DNS Private Resolver is a managed service: its inbound endpoint gives on-premises DNS an IP to forward Azure queries to, and an outbound endpoint with a forwarding ruleset sends corp.contoso.com queries to on-premises DNS. The 168.63.129.16 address is reachable only from inside Azure, not from on-premises. DNS VMs work but add the management the company wants to avoid, and Azure public DNS does not support split-horizon views.
Question 10 of 10, Design infrastructure solutions
An AKS platform team wants a managed layer-7 load balancer for its clusters that supports traffic splitting and header-based routing, is configured from Kubernetes Gateway API resources, and runs outside the cluster so it does not consume node resources. What should you recommend?
- AThe NGINX ingress controller deployed as pods on a dedicated system node pool in the cluster
- BAzure Front Door Standard with each AKS service exposed through a public IP
- CAn internal Standard Load Balancer with a separate service per traffic split
- DApplication Gateway for Containers managed by the ALB Controller in the cluster Correct answer
Answer and explanation: Application Gateway for Containers is an Azure-managed layer-7 load balancer that runs outside the cluster, supports traffic splitting and header-based routing, and is configured through the Gateway API by the ALB Controller. NGINX runs inside the cluster on nodes. Front Door is a global entry point that does not read Gateway API resources, and a layer-4 load balancer cannot route on headers.