Microsoft AZ-104 Exam Dumps, Practice Test Questions

Zone selection affects those VM resources, not automatically the database or external dependencies. The single-zone database is still a failure point.

A zone-redundant frontend can keep routing to healthy web instances after a zone loss, but it cannot make a database that exists only in zone 1 available. The required stateful backend still fails and interrupts the application; that dependency needs a cross-zone resilience design.

Protecting the web VM helps recover a stateless tier, but the application still requires its stateful database in the failed zone. Without a resilient or recoverable database dependency and tested failover, the service remains unavailable despite other recovery features.

Read-only attachment behavior does not override a managed disk’s zonal compatibility requirements. This disk is pinned to another availability zone, so the platform cannot attach it to the stated zone-2 VM merely by withholding write permission. Create or select a disk whose placement and type are supported for the target VM.

A managed disk snapshot is a recovery source, not an attachable managed disk that can be presented to a running VM as-is. To move disk content across zones, first create an eligible new managed disk from the snapshot in a supported destination zone and then attach that disk. Directly attaching the snapshot would not satisfy the compute disk resource requirement.

A resource group is a management scope and can contain Azure resources in different physical locations. Moving a managed disk between resource groups does not change its zone assignment. The attachment failure is due to the disk and VM having incompatible zonal placement; use an eligible disk in the VM zone or another supported design.

Non-zonal placement has its own availability behavior and does not force common physical facilities. Removing zone requests would not prove the intended relationship between subscriptions.

Alert severity controls event priority, not physical hardware placement. Identical alert rules do not establish which datacenter hosts the workloads.

Tags do not control Azure's physical datacenter placement. A matching metadata value has no ability to align zone mappings across subscription boundaries.

Shared Premium SSD disks are data disks, not a method to boot multiple VMs from one OS disk. Two independent cluster nodes require their own boot disks while they coordinate access to a common data LUN. Making it an OS disk would violate the shared-disk design.

Availability sets address placement across fault and update domains. They do not change a disk's sharing limit; the supported Premium SSD data disk must have maxShares set appropriately and cluster-aware coordination enabled before multiple VMs can attach it.

Azure doesn't infer multiattach permission from matching VM names or a shared subnet. A normal single-attach disk rejects incompatible additional attachments. The disk must be a supported shared SKU with maxShares configured appropriately, and the guest cluster must handle concurrent writers safely.

Disk caching controls guest I/O caching behavior and performance. It does not choose a managed disk's infrastructure fault domain or substitute for availability-set placement.

Tags annotate Azure resources but don't control hardware placement. They cannot align or separate managed-disk storage fault domains for an availability-set workload.

Proximity placement groups seek low-latency co-location, not the same resilience mechanism as fault-domain distribution. Using one does not establish the desired separation of fault risks.

Deleting the group removes the existing colocation intent, and separating VMs into different regions would ordinarily increase physical distance rather than restore low-latency alignment. The maintenance event has disrupted alignment inside an existing group, not demonstrated a need for cross-region placement. Recovery should address the placement allocation of affected resources.

Changing the proximity placement group association for a running VM is not an in-place host realignment operation; changing placement association requires stopping and deallocating an affected VM as appropriate. Leaving the affected VM on the same allocation does not repair the reported Not aligned colocation state. Use the documented deallocate/start realignment flow.

Size intent helps Azure plan suitable deployment capacity, but editing intent does not relocate the already-running affected VM back into the group’s latency envelope. The observed issue is current physical colocation after host maintenance; deallocating and restarting the affected VM is the direct documented step, with larger-group remediation only if needed.

The update-domain count is configured on the availability set and doesn't expand dynamically because one more VM has joined. Azure reuses existing domains. Assuming an automatic sixth domain could cause the team to misunderstand which VMs may restart together in planned host maintenance.

Fault domains protect against shared physical fault boundaries, while update domains stage planned maintenance restarts. Azure normally processes one update domain at a time and allows time for recovery before another. A failure to distinguish these two placements leads to an incorrect simultaneous-restart assumption.

An availability set is a VM placement construct within a region, not automatic cross-region or cross-zone migration. Adding a sixth VM doesn't send it to another region. For regional or zonal resiliency, separately design workload replicas and zone or region placement.

Update domains stagger planned platform maintenance for VMs in an availability set. Their count has no automatic relationship to a Standard Load Balancer’s address-pool membership. Raising that count cannot enroll Web2 as a backend; the relevant NIC IP configuration or address must be associated with the pool.

An inbound NAT rule forwards a specified frontend protocol and port to one particular backend address/port. It does not make an omitted VM participate in the load-balancing rule’s shared backend pool. The request distribution path requires Web2 to be included in that pool rather than creating a separate one-to-one NAT mapping.

Disabling a required health probe does not change which backend addresses are configured and removes a necessary signal for healthy traffic routing. The scenario explicitly states Web2 is not in the pool, so probe changes cannot repair the missing association and may make otherwise healthy traffic handling less reliable.

Availability-set fault-domain cardinality is configured when the set is created and is not expanded automatically each time a VM is added. With two configured fault domains, Azure can distribute replicas across those two power-and-network failure groups but cannot invent a third independent group. Increasing VM count does not change the count of distinct fault domains.

Availability sets use fault and update domains inside a regional deployment and do not automatically create or guarantee zone-spread for the third VM. Availability zones are a separate placement architecture. The question explicitly specifies a two-fault-domain set, so adding a third replica cannot grant an unstated third-zone guarantee.

Azure availability sets can contain more VMs than their configured number of fault domains. The platform distributes those VMs over the available failure groups rather than enforcing a one-VM-per-fault-domain maximum. A third VM may be provisioned, but the architecture cannot guarantee three independently isolated physical fault domains.

A private IP address describes how an instance is addressed. It does not monitor utilization or change the number of running compute instances.

Fixed capacity may cover the busiest period but wastes resources when usage falls, and it does not automatically respond to sustained metric changes.

An alert can notify staff of load, but it does not itself change instance count. The objective is automated capacity adjustment as demand changes.

An instance may be running while its application returns errors. Instance count alone cannot identify application-specific failures.

ProvisioningState reports the management-plane deployment state, not the health of the running HTTP API. An instance can finish provisioning while its process is down or returns failing responses. Automatic upgrades and repairs require a supported application health signal or probe aligned with the actual backend behavior, not merely a provisioning-success value.

Power or instance provisioning state does not establish that an HTTP listener answers requests correctly. A VM can be running while its API fails health checks. VMSS upgrade and repair decisions need a supported application health signal such as the Application Health extension rather than treating VM running state as application health.