Cost Insights
Cost Insights
Warning: The costs shown in Cost Insights are estimates and may differ from your actual Azure invoices. Final billing is determined solely by Microsoft at the time of invoicing and may depend on such factors as currency exchange rates, discounts, or adjustments that are not reflected in this information.
This chapter focuses solely on the session host servers in each selected host pool. Costs associated with Parallels RAS infrastructure components, such as the Secure Gateway and Connection Broker, as well as Azure infrastructure components like the Azure Load Balancer, are currently not included in the calculations below.
System requirements
Parallels RAS Cost Insights can be installed on a server running one of the following Windows Server versions:
Windows Server 2025
Windows Server 2022
Windows Server 2019
Windows Server 2016
Other requirements include:
Parallels RAS Cost Insights requires Parallels RAS Reporting to be configured first.
Parallels RAS Cost Insights requires Azure Virtual Desktop to be configured first, including an Azure Virtual Desktop Provider.
Parallels RAS Cost Insights needs to be installed on the same machine as Parallels RAS Reporting and Microsoft SQL Server Reporting Services.
Parallels RAS Cost Insights required an App Registration to be configured in Azure. This App Registration can be the same as configured on the provider, and needs at least reader permission on all Resource Groups where Azure Virtual Desktop resources are located, and Billing Reader permissions on the Azure subscription.
Accessing Parallels RAS Cost Insights from within the Parallels RAS console requires the WebView2 runtime. This is installed by default on most operating systems, except for Windows Server 2022. If the WebView2 runtime is missing, install it manually or follow the steps provided upon initially opening Cost Insights, which guide you through installing the Parallels RAS Browser Plugin.
Install Parallels RAS Cost Insights
To install Parallels RAS Cost Insights:
Log in to the server where Parallels RAS Reporting is installed. Make sure to use an account with administrative privileges (AD).
Double-click the
RASInstaller.msifile to launch the Parallels RAS installation wizard. If you see a message that begins with "This version of Parallels RAS is only for testing purposes", it means that it's not an official build and should not be used in a production environment.Follow the onscreen instructions.
Proceed to the Select Installation Type page and select Custom.
Select the component Cost Insights only, and unselect all other components.
Click Next.
On the Firewall Settings page, select Automatically add firewall rules to configure the firewall on this computer for Parallels RAS to work properly. See Port Reference for details.
Click Next and then click Install. Wait for the installation to finish and click Finish.
If you are upgrading your RAS installation, it is recommended to reboot all servers where components are upgraded.
Configure Parallels RAS Cost Insights
To configure Parallels RAS Cost Insights:
Log in to the Parallels RAS Console.
Select the Administration category and click the Reporting & Cost Insights tab in the right pane.
In the Reporting tab, select the Enable RAS Cost Insights option and click Apply.
Select the Farm category and click the Providers tab in the right pane.
Open the provider you want to collect costs for and switch to the Advanced tab.
Select Collect costs for this provider.
Optionally, modify the app registration to use, or use the same app registration as configured on the provider. Consult the requirements to confirm the specified app registration has the required permissions.
Click Apply.
Select the Administration category and click the Reporting & Cost Insights tab in the right pane.
To test the RAS Cost Insights connection, click the Test connection button.
Managing Parallels RAS Cost Insights
The Cost Insights page allows you to review and analyze the estimated Azure infrastructure costs associated with your session host pools. From this page, you can monitor both current and historical trends, evaluate cost efficiency, and identify opportunities for optimization using Parallels RAS features.
Log in to the Parallels RAS Console.
Select the Cost Insights category.
Use the Time Range dropdown to choose the reporting interval (e.g., This week, Last month, or a custom date range). This determines which time window the cost calculations are based on and affects all metrics displayed.
Use the Selected Host Pool dropdown to focus the analysis on a specific host pool. If multiple host pools exist, you can switch between them to compare costs and utilization.
The Virtual Machine Type field displays the VM size(s) used by the selected host pool (for example, Standard_B2s). This is for reference only and cannot be used as a filter. It helps you understand the cost breakdown in the context of the VM instance type(s) currently in use.
Click on the information icons next to each metric (e.g., Host Pool Cost, Compute Cost, Parallels RAS Current Savings) to see detailed definitions.
Compare Parallels RAS Current Savings with Parallels RAS Potential Savings to understand the actual and possible benefits of enabling optimization features. This helps you determine whether enabling disk storage optimization or autoscaling can further reduce costs.
Use Total Cost per User to benchmark the cost efficiency of your host pool. Combine this with the # of Active Users to see how effectively your host pool resources are being utilized.
Scroll to the Historical Analysis graphs to view cost and cost-per-user trends over time. These visualizations help you correlate peaks in costs with usage patterns, scaling events, or configuration changes.
Definitions and calculations
General
All costs shown in the Cost Insights Dashboard are based on the UTC time zone. This is because both the Azure Cost Management API and the Azure Portal operate exclusively in UTC, and all cost data in Azure is emitted and stored in UTC.
For example, when selecting “Yesterday” as the time range, the dashboard displays costs from 00:00:00 to 23:59:59 UTC for that calendar day.
The costs shown in Cost Insights are estimates and may differ from your actual Azure invoices. Final billing is determined solely by Microsoft at the time of invoicing and may depend on such factors as currency exchange rates, discounts, or adjustments that are not reflected in this information.
Host Pool Cost
The total Azure cost for all session host virtual machines that are part of the selected host pool(s) within the specified time range. This includes three major components: Compute, OS Disk Storage, and Network. Compute represents the VM size and runtime costs, OS Disk Storage represents the managed disks attached to each session host, and Network represents outbound and inter-region traffic costs. This metric provides you with a consolidated view of the overall infrastructure cost directly associated with operating the session hosts in the selected host pool(s).
Calculation Details
Host Pool Cost = Σ (Compute cost of each session host) + Σ (OS disk cost of each session host) + Σ (Network cost of each session host)
Compute Cost
The compute cost of all Azure session host virtual machines that belong to the selected host pool(s) within the specified time range. The compute cost is based on the VM size (vCPU, memory, and GPU, if applicable) and the number of hours each session host runs. Costs are calculated according to the Azure pricing model for the chosen VM sizes. This metric reflects the single most significant portion of host pool cost and is directly influenced by autoscaling configurations, VM uptime, and workload demand.
Calculation Formula:
Compute Cost = Σ (Compute cost of each session host)
Storage Cost
The operating system (OS) disk storage cost for all session host virtual machines in the selected host pool(s) within the specified time range. This is based on the size and type of managed disks (Standard HDD, Standard SSD, or Premium SSD) provisioned for each session host. Storage cost reflects the recurring per-GB/month rate charged by Azure for disk allocation, regardless of how much of the disk space is actively used. Temporary disks, data disks, and ephemeral OS disks (if configured) are excluded from the calculation.
Calculation Formula:
Storage Cost = Σ (OS disk cost of each session host)
Network Cost
The network cost associated with all session host virtual machines in the selected host pool(s) during the specified time range. This cost is derived from Azure bandwidth charges, specifically for data transfer out across continents and inter-region ingress/egress between Azure regions. Intra-region and inbound data transfer are typically not billed, but outbound and cross-region transfers can add up significantly depending on workload patterns. This metric consolidates those Azure network costs attributable to session host activity.
Calculation Formula:
Network Cost = Σ (Network cost of each session host)
Parallels RAS Current Savings
The realized cost savings that have been achieved through the configuration of Parallels RAS optimization features in the selected host pool(s) during the specified time range. This currently includes two mechanisms:
Disk Storage Optimization: savings by using optimized disks when VMs are stopped and deallocated.
Autoscaling: savings from powering VMs on or off based on actual demand.
This metric represents actual avoided costs compared to what would have been incurred if all session hosts had run continuously without optimization.
Calculation Formula:
Disk Storage Optimization Savings (if enabled on the RAS provider):
Σ for each session host:
((Deallocated hours) x (unoptimized disk price per hour)) -
((Deallocated hours) x (optimized disk price per hour))
Autoscaling Savings (if enabled):
Σ for each session host:
(Deallocated hours) x (compute price per hour for VM type)
Parallels RAS Potential Savings
The additional cost savings that could have been achieved if Parallels RAS optimization features (disk storage optimization and autoscaling) had been fully enabled and utilized for the selected host pool(s) and time range. Unlike Current Savings, which represents realized reductions, this metric highlights the potential for further savings. It is a projection based on observed usage patterns and assumes full adoption of RAS scaling and storage policies.
Calculation Formula:
Potential Disk Storage Savings (if optimization is disabled):
Toff = time the host was deallocated C_unopt = unoptimized disk cost per hour C_opt = optimized disk cost per hour Potential Storage Savings = Σ for each session host: Toff x (C_unopt – C_opt)
Potential Autoscaling Savings (if power management is disabled):
Σ for each session host:
(Hours the VM was running with no active sessions) x (compute price per hour for that VM type)
Total Cost per User
The host pool cost, divided by the number of active users during the selected time range. This provides an average cost-per-user metric that helps you measure the efficiency of your deployment. By normalizing total infrastructure cost against peak user load, this metric offers a benchmark for cost-effectiveness and helps you compare host pool performance against other pools or industry baselines.
Calculation Formula:
Total Cost per User = Host Pool Cost ÷ Maximum # of Active Users
Maximum # of Session Hosts
The highest number of session host virtual machines that were simultaneously running across all selected host pool(s) during the specified time range. This metric reflects the historical peak scale of the environment and helps you analyze capacity requirements, evaluate autoscaling performance, and plan for future demand.
Calculation Formula:
Maximum # of Session Hosts = Max (Count of session hosts in "Running" state over the selected time range)
# of Active Users
The number of unique users who were active over the selected time range for all selected host pool(s). Active means the user had at least one active session selected in the time range for all selected host pool(s).
Calculation Formula:
Maximum # of Active Users = Max (Count of users with active sessions over the selected time range)
Historical Analysis: Cost per User
A time-series graph showing how the cost per user has varied over the selected time range for all selected host pool(s). Costs are normalized against active user counts for each time interval (e.g., hourly, daily, or weekly). This allows you to visualize trends in cost efficiency, identify peak inefficiencies, and assess the impact of scaling or optimization changes over time.
Historical Analysis: Host Pool Cost
A time-series graph showing the total host pool cost (compute, OS disk storage, and network) over the selected time range for all selected host pool(s). This allows you to monitor historical cost patterns, correlate costs with user activity or scaling events, and identify opportunities for cost optimization.
Last updated
Was this helpful?