AZ-140 Domain 4: Monitor and maintain an Azure Virtual Desktop infrastructure (10–15%)

Overview

Monitor and maintain an Azure Virtual Desktop infrastructure has two functional groups and 11 skills. Monitoring and management covers log collection to Log Analytics, Azure Monitor, customizing the Azure Virtual Desktop Insights workbooks, optimizing session host capacity and performance, autoscaling host pools, and managing active sessions and application groups.

The second group covers update strategy for session hosts, disaster recovery planning, multi-region design, backup strategy, and backup and restore for FSLogix profiles, personal desktops and images. Although weighted at only 10–15%, this domain borrows heavily from the others: you cannot plan recovery without knowing where profiles, images and host pool metadata live. Questions favour operational judgement, such as which scaling plan phase or recovery method meets a stated target.

Why it matters on the exam

Operations is where Azure Virtual Desktop costs are won or lost. A scaling plan that never deallocates hosts doubles the compute bill, and a missing FSLogix backup turns a storage incident into lost user data. Microsoft weights this domain at 10–15%, the smallest share, but the questions are often easy points for candidates who have run a deployment, and costly for those who have not. You need to know which tables appear in Log Analytics, what Insights needs before it shows data, how pooled and personal scaling plans differ, and which Azure service protects each component. Because the material builds on the other three domains, review it after them and use it to tie your whole study plan together.

How it's tested

Domain 4 questions describe an operational goal such as reducing cost overnight, finding the cause of failed connections, or meeting a recovery time target, then offer services or settings. Most distractors are real Azure features that work for general VMs but miss an Azure Virtual Desktop specific need, like scaling plans or Insights data collection.

Find the data

A troubleshooting or reporting need is given, and you choose diagnostic settings, Log Analytics tables or the Insights workbook to use.
Scale for cost

Usage patterns are described, and you configure scaling plan schedules, capacity thresholds, minimum hosts or personal desktop power actions.
Recover the service

A failure or compliance scenario is stated, and you choose backup, replication or a secondary-region host pool approach.

Reasoning it through

A pooled host pool runs ten hosts all day, yet usage drops sharply after 6 p.m. The team wants hosts shut down automatically while users stay connected. A scaling plan with a ramp-down phase starting at 6 p.m., a capacity threshold and a minimum percentage of hosts deallocates idle hosts gradually. A scheduled runbook could work, but the built-in scaling plan is the supported answer.

Functional groups 2

Official functional groups from the Microsoft outline, in outline order.

4.1 Monitor and manage Azure Virtual Desktop servicesLog collection, Azure Monitor, Insights workbooks, capacity optimization, autoscaling, and session and application group management.

What it means: This group covers sending Azure Virtual Desktop diagnostics and session host performance data to Log Analytics, monitoring with Azure Monitor, customizing Azure Virtual Desktop Insights workbooks, optimizing capacity and performance, implementing autoscale scaling plans, and managing active sessions and application groups.
What to understand: Diagnostic settings on host pools, workspaces and application groups write tables such as WVDConnections, WVDErrors and WVDAgentHealthStatus to Log Analytics. Insights also needs a data collection rule for performance counters and event logs from session hosts through the Azure Monitor Agent. Scaling plans use ramp-up, peak, ramp-down and off-peak phases for pooled pools, and power actions on disconnect or sign-out for personal pools. They require the Desktop Virtualization Power On Off Contributor role for the service principal.
How it shows up: Questions ask what data source, workbook, scaling setting or session action solves an operational problem.
Common confusion: Expecting Insights to show session host performance without a data collection rule and the Azure Monitor Agent.

4.2 Plan and implement updates, backups, and disaster recoveryUpdate strategy, disaster recovery, multi-region design, backup strategy, and backup and restore for profiles, personal desktops and images.

What it means: This group covers recommending how session hosts are patched, planning disaster recovery and multi-region deployments, designing a backup strategy, and configuring backup and restore for FSLogix profiles, personal desktops and images.
What to understand: Pooled hosts are best updated by building a new image version and replacing hosts, while personal desktops are usually patched in place with Azure Update Manager or Intune. Disaster recovery usually means a host pool in a secondary region with replicated images, replicated or Cloud Cache protected FSLogix storage, and a plan for host pool metadata. Azure Backup protects Azure Files shares and personal desktop VMs, Azure Site Recovery can replicate personal VMs, and Compute Gallery replicas protect images.
How it shows up: Expect scenarios with recovery targets or patching constraints, asking which update or recovery method meets them.
Common confusion: Backing up pooled session host VMs individually; they are disposable, so protect images and profile storage instead.

High-yield concepts

The terms, rules and formulas this domain turns into questions most often.

Diagnostic settings

Send AVD service logs to Log Analytics, storage or Event Hubs.

Remember
Configure on host pools, workspaces and application groups.

Tested as
Troubleshooting items.

Trap
Enabling them only on session host VMs.
WVDConnections and WVDErrors

Log Analytics tables for connections and failures.

Remember
Query them to trace failed sign-ins.

Tested as
Log analysis items.

Trap
Searching VM performance tables for connection errors.
Azure Virtual Desktop Insights

Azure Monitor workbook for AVD health and usage.

Remember
Needs diagnostics plus a data collection rule.

Tested as
Monitoring setup items.

Trap
Assuming it works with no configuration.
Data collection rule

Defines counters and logs collected by the Azure Monitor Agent.

Remember
Insights uses a specific set of counters.

Tested as
Monitoring setup items.

Trap
Relying on the legacy Log Analytics agent.
Scaling plan (pooled)

Schedule-based autoscale with four phases.

Remember
Capacity threshold and minimum hosts control decisions.

Tested as
Cost optimization items.

Trap
Setting minimum hosts to 100% off-peak.
Scaling plan (personal)

Starts and stops personal desktops by schedule and user activity.

Remember
Can hibernate, deallocate or do nothing on disconnect.

Tested as
Personal pool cost items.

Trap
Using pooled settings for personal pools.
Power On Off Contributor

Role needed by the service principal for autoscale.

Remember
Assign at subscription or resource group scope.

Tested as
Autoscale troubleshooting.

Trap
Assigning only the Power On role.
Session management actions

Send message, disconnect or sign out users.

Remember
Combine with drain mode before maintenance.

Tested as
Operations items.

Trap
Rebooting hosts with users signed in.
Image-based patching

Update the image, then redeploy pooled hosts.

Remember
Keeps hosts consistent.

Tested as
Update strategy items.

Trap
Patching each pooled host by hand.
Azure Backup for Azure Files

Snapshot-based protection for FSLogix shares.

Remember
Restores whole shares or single files.

Tested as
Profile recovery items.

Trap
Relying on host VM backups for profiles.
Azure Site Recovery

Replicates VMs to a secondary region.

Remember
Fits personal desktops, not disposable pooled hosts.

Tested as
DR design items.

Trap
Using it for pooled hosts.
Host pool metadata location

Region where the service stores host pool and workspace data.

Remember
Can differ from session host region.

Tested as
Multi-region planning.

Trap
Assuming metadata follows the VMs.

Side by side

Concepts candidates mix up, compared on the point that decides the answer.

Monitoring tools

Concept	Meaning	Key difference	Common trap
Diagnostic settings	Export service logs to a destination.	Raw data source for queries.	Expected to show dashboards on their own.
Azure Virtual Desktop Insights	Prebuilt workbook views of health and usage.	Visualizes diagnostics and host counters.	Missing its data collection rule.
Azure Monitor alerts	Notify on metric or log conditions.	Proactive notification, not analysis.	Used where historical analysis is required.

Scaling approaches

Concept	Meaning	Key difference	Common trap
Pooled scaling plan	Phase-based start and stop of shared hosts.	Uses load and capacity thresholds.	Applied to personal pools.
Personal scaling plan	Power actions per personal desktop.	Reacts to user connect, disconnect and sign-out.	Expected to load balance users.
Start VM on Connect	Starts a deallocated host when a user connects.	Reactive, complements scaling plans.	Used alone to shut hosts down.

Protecting each component

Concept	Meaning	Key difference	Common trap
FSLogix profiles	User data on SMB storage.	Azure Backup, storage replication or Cloud Cache.	Backed up through pooled VMs.
Personal desktops	Stateful single-user VMs.	Azure Backup or Azure Site Recovery.	Treated as disposable.
Images	Golden image versions.	Compute Gallery replicas across regions.	Stored in one region only.

Exam traps

Each trap is an answer that looks right. Learn the clue that should change your mind.

Runbook over scaling plan

What candidates pick
Write an Azure Automation runbook to stop hosts.

Why it looks right
It was the older documented approach.

Why it's wrong
Native scaling plans now provide schedule and capacity based autoscale without custom scripts.

Clue to watch for Requirement asks for a built-in or supported feature.
Insights with no host data

What candidates pick
Reinstall the session host agent.

Why it looks right
Agent issues are a common cause of trouble.

Why it's wrong
Host performance data requires a data collection rule and the Azure Monitor Agent, not an agent reinstall.

Clue to watch for Workbook shows service data but no host counters.
Backing up pooled hosts

What candidates pick
Enable Azure Backup on every pooled VM.

Why it looks right
Backing everything up feels safest.

Why it's wrong
Pooled hosts hold no user state; protect images and FSLogix storage and rebuild hosts instead.

Clue to watch for Hosts are pooled and non-persistent.
Site Recovery for everything

What candidates pick
Replicate all session hosts with Azure Site Recovery.

Why it looks right
It is Azure's VM disaster recovery service.

Why it's wrong
For pooled hosts a secondary host pool built from replicated images is simpler and cheaper.

Clue to watch for The scenario mixes pooled and personal hosts.
Rebooting during the day

What candidates pick
Restart hosts immediately to apply updates.

Why it looks right
Updates need a reboot.

Why it's wrong
Drain mode, user messages and a maintenance window avoid disconnecting active users.

Clue to watch for Users are still signed in.

Answer framework

Use these steps for monitoring, scaling, patching and recovery questions in AZ-140.

Name the operational goal clearly: troubleshooting data, cost reduction, patching, or recovery from a regional or storage failure, and note any time or budget limit.
For monitoring, confirm diagnostic settings on service objects and a data collection rule on hosts before choosing Insights or log queries.
For cost, choose a pooled or personal scaling plan, set phases and thresholds, and verify the service principal role assignment.
For updates, prefer new image versions and host replacement for pooled hosts, and in-place patching tools for personal desktops.
For recovery, list each component separately: images, FSLogix storage, personal VMs and host pool metadata, then protect each appropriately.

Practice question

A pooled host pool uses Azure Virtual Desktop Insights. The workbook shows connection data but no CPU or memory counters from session hosts. Diagnostic settings are enabled on the host pool and workspace. What should the administrator configure next?

An Azure Monitor alert rule for high CPU A data collection rule with the Azure Monitor Agent on the session hosts A new registration token for each session host A scaling plan with a ramp-up phase

Show the answer

Answer B. Insights reads service diagnostics from the host pool and workspace, which are already working, and host performance counters gathered by the Azure Monitor Agent through a data collection rule. Without that rule, no CPU or memory data reaches Log Analytics. Alert rules only notify, a registration token affects host pool membership, and a scaling plan changes capacity rather than monitoring data.

Want more like this? The AZ-140 practice test has questions for every domain.

Checklist

0 of 11

Tick each one you can do without notes. Your ticks stay in this browser.

Study strategy

Plan about 10 to 14 hours, ideally after you have studied the other three domains. Begin by enabling diagnostic settings on your lab host pool and workspace, then set up Azure Virtual Desktop Insights so you see which parts populate immediately and which wait for the data collection rule. Run a few simple queries against WVDConnections and WVDErrors after deliberately causing a failed connection.

Next, create a pooled scaling plan, assign it, and observe the ramp-down behaviour, then build a personal scaling plan and compare its options. Practice draining a host, messaging users and signing them out, and note which built-in role a help desk operator would need for those actions. Finish by drawing a recovery diagram that lists each component, where it lives, and the Azure service that protects it. Review that diagram with a timed set of practice questions, because recovery scenarios often combine facts from storage, images and identity covered elsewhere in AZ-140.

FAQ

How much of AZ-140 is Monitor and maintain an Azure Virtual Desktop infrastructure?

Monitor and maintain an Azure Virtual Desktop infrastructure is weighted at 10–15%, the smallest of the four domains. That works out to roughly 5 to 8 questions as an editorial estimate, since Microsoft publishes no per-domain counts.

Do I still need Azure Automation runbooks for scaling?

No. Scaling plans are the built-in autoscale feature for pooled and personal host pools, and they are what current exam scenarios expect.

What does Azure Virtual Desktop Insights need to work?

It needs diagnostic settings on Azure Virtual Desktop objects sending data to a Log Analytics workspace, plus a data collection rule so the Azure Monitor Agent collects session host counters and events.

Should I back up pooled session host VMs?

Usually not. Pooled hosts hold no user data and can be rebuilt from an image. Protect the image versions and FSLogix storage instead.

How is disaster recovery different for personal desktops?

Personal desktops hold user state, so they need VM-level protection such as Azure Backup or Azure Site Recovery to another region, in addition to profile and image protection.

Can Azure Virtual Desktop metadata and session hosts be in different regions?

Yes. The service stores host pool and workspace metadata in a chosen location, while session hosts can run in any supported Azure region, which matters for multi-region planning.

Related domains

Other Cloud Computing certifications

Where the same subject shows up in other exams. Cards marked "Same subject" open the matching domain.

AZ-140 Domain 4: Monitor and maintain an Azure Virtual Desktop infrastructure

Overview

How it's tested

Find the data

Scale for cost

Recover the service

Functional groups 2

High-yield concepts

Diagnostic settings

WVDConnections and WVDErrors

Azure Virtual Desktop Insights

Data collection rule

Scaling plan (pooled)

Scaling plan (personal)

Power On Off Contributor

Session management actions

Image-based patching

Azure Backup for Azure Files

Azure Site Recovery

Host pool metadata location

Side by side

Monitoring tools

Scaling approaches

Protecting each component

Exam traps

Runbook over scaling plan

Insights with no host data

Backing up pooled hosts

Site Recovery for everything

Rebooting during the day

Answer framework

Practice question

Checklist

Study strategy

FAQ

Other Cloud Computing certifications

Test your Domain 4 knowledge

Keep going