The Performance Monitoring & Diagnosis page gives you a real-time view of resource utilization across your cloud computers — CPU, memory, disk, and a composite Load Score. Use it to spot overloaded machines, compare usage across your fleet, and restart struggling instances before end users notice.
Metric collection requires the CloudMonitor agent (argusagent service) to be running on each cloud computer. The agent is disabled by default. If metrics are missing or alerts are not firing, see Why aren't alerts firing?.Prerequisites
Before you begin, ensure that you have:
Permission to view or manage cloud computers. For details, see Overview
View metrics
The Performance Monitoring & Diagnosis page has three tabs, each covering a different scope:
| Tab | What it shows |
|---|---|
| Cloud Computer | Individual cloud computers assigned to users |
| Shares | Shared cloud computers (multi-session) |
| Top 50 Resources | The 50 cloud computers with the highest resource consumption |
Each tab lets you view metric graphs — including Load Score, CPU usage, memory usage, and disk-related parameters — over a selected time interval.
Load Score is not available on the Performance Monitoring tab of the cloud computer details page. That tab shows CPU usage, memory usage, and disk-related parameters only.
From the Performance Monitoring & Diagnosis page
Log on to the Elastic Desktop Service Enterprise console.
In the left navigation pane, choose Monitoring & Alerts > Performance Monitoring & Diagnosis.
In the top navigation bar, select a region.
Click the Cloud Computer, Shares, or Top 50 Resources tab.
In the Actions column of the target cloud computer:
On the Cloud Computer tab, click Performance Monitoring.
On the Shares or Top 50 Resources tab, click Graphs.
On the Performance Monitoring or Monitoring Details page, select a preset time interval or specify a custom range to view the metric graphs.
From the cloud computer details page
This method applies to individual cloud computers only. Shared cloud computers are not supported.
Log on to the Elastic Desktop Service Enterprise console.
In the left navigation pane, choose Resources > Cloud Computers.
In the top navigation bar, select a region.
On the Cloud Computers page, click the ID of the target cloud computer.
On the Performance Monitoring tab, select a time interval to view CPU usage, memory usage, and disk-related parameters for that machine.
Send messages
Send a message to notify end users of a pending restart so they can save open files before the session ends.
Before you send messages, the following requirements must be met:
The cloud computer must run Windows and use the Adaptive Streaming Protocol (ASP).
The cloud computer image must be a system image V 1.0.0 or later, or a custom image based on a system image V 1.0.0 or later.
End users must be on the Windows client or macOS client V 5.0.0 or later.
In the left navigation pane, choose Monitoring & Alerts > Performance Monitoring & Diagnosis.
In the top navigation bar, select a region.
Click the Cloud Computer, Shares, or Top 50 Resources tab. Select the cloud computers you want to message, then click Send Message at the bottom of the page.
In the Send Message dialog, fill in Subject and Messages, then click Confirm.
The message appears on the end user's screen.
Restart a cloud computer
When a cloud computer shows abnormal metrics, restart it to clear the issue.
Before restarting, send a message to the cloud computer to give connected end users time to save their work. See Send messages.
In the left navigation pane, choose Monitoring & Alerts > Performance Monitoring & Diagnosis.
In the top navigation bar, select a region.
Click the Cloud Computer, Shares, or Top 50 Resources tab and find the cloud computer you want to restart.
In the Actions column, click Restart, then click Confirm in the dialog.
To restart multiple cloud computers at once, select them and click Restart at the bottom of the page.
Export cloud computer resource data
Export resource data as an XLSX file to analyze usage offline or share with your team. The export includes cloud computer ID and name, specifications/system, purchase method, CPU usage, memory usage, system disk usage, data disk usage, and GPU memory usage.
In the left navigation pane, choose Monitoring & Alerts > Performance Monitoring & Diagnosis.
In the top navigation bar, select a region.
Click the Cloud Computer, Shares, or Top 50 Resources tab.
Select the cloud computers you want to export (required for Export Selected), then click the
icon in the upper-right corner.In the dialog, select an export scope:
Export Selected: exports the cloud computers you selected in step 4.
Export All: exports all cloud computers on the current tab.
Click Confirm.
In the export task notification that appears in the upper-right corner, click Viewing export records.
If the notification disappears before you click it, open the sidebar on the right side of the console and click the export records icon to find past exports.
In Export Records, wait for the Status to change to Export Success, then click Download in the Actions column.
FAQ
Why aren't alerts firing even though I have an alert rule configured?
The CloudMonitor agent that collects metric data on cloud computers is disabled by default. Enable it as follows:
On your cloud computer, press Win+R, type
services.mscin the Run window, and press Enter.In the Services window, find argusagent service, right-click it, and select Properties.
On the General tab, set Startup type to Automatic. If the service isn't running, click Start under Service status, then click OK.