Troubleshooting and Resolving Performance Limitations in Azure Virtual Desktop (AVD) Environments: A Comprehensive Guide

 

Providing a seamless user experience requires Azure Virtual Desktop (AVD) environments to deliver optimal performance. However, performance limitations in components such as the CPU, RAM, disc, and network can be caused by a number of factors. This exhaustive guide will delve into troubleshooting performance limitations in Azure Virtual Desktop and identifying each component, defining potential issues, and providing detailed troubleshooting and resolution procedures to ensure a high-performance AVD environment. To learn more about AVD visit our previous blog post here.

 

CPU-Related Performance Limitations:

In AVD environments, CPU performance is crucial, and addressing CPU-related issues is essential for sustaining a responsive and efficient system. Here is a comprehensive approach to diagnosing and resolving CPU-related performance limitations:

 

Problem: High CPU Usage or Spikes

High CPU utilisation or frequent surges can result in sluggish performance, delay application responsiveness, and aggravate users. To resolve this matter, please follow these steps:

 

Step 1: Identify Processes Responsible for High CPU Usage

Utilise the performance tab of Task Manager to monitor CPU utilisation on the session host VM. Determine whether CPU utilisation is consistently low or moderate or frequently surges to high levels. If CPU utilisation is always minimal, you should concentrate your troubleshooting efforts on other system components. However, if CPU usage increases frequently, further analysis is required.

Step 2: Eliminate or Improve violating Processes

Navigate to the details pane of Task Manager and sort the processes by CPU utilisation. Examine the processes at the top of the sorted list for any that are unanticipated or should not be utilising a significant amount of CPU resources. Examine these processes to determine their function, the user context in which they execute, and the reason for their high CPU usage.

If a process is deemed unnecessary or detrimental to performance, it should be terminated using Task Manager. Analyse the resource requirements of critical or optimisation-required processes and investigate potential solutions. Consider involving application developers in the performance optimisation process or in the search for alternate solutions.

Step 3: Boost Capacity or Reduce Load

Consider increasing the size of the virtual machine (VM) to provide more CPU capacity if all processes are operating as anticipated and are not consistently consuming excessive amounts of CPU. Alternately, distribute users across more session host VMs to effectively distribute the CPU burden and improve overall performance.

 

RAM-Related Performance Limitations:

Insufficient RAM can slow down AVD environments and impact their overall performance. Here is a comprehensive method for diagnosing and resolving RAM-related performance limitations:

 

Problem: High RAM Usage and Critical Errors

When RAM usage is high and severe faults occur frequently, the system suffers from a RAM shortage. This may result in a decrease in performance due to the necessity of retrieving data from a disk rather than RAM. To resolve this matter, please follow these steps:

 

Step 1: Monitor Hard Faults and RAM Usage.

Start Windows Resource Monitor from the performance pane of Task Manager or by typing “resmon” in the Run box. Pay close attention to the Hard Faults per second counter on the Memory pane. Little to no activity indicates that RAM is not limiting the system’s performance, whereas bursts or continual hard fault activity indicate a RAM-related performance issue.

Step 2: Identify Contributing Processes for Hard Faults

Sort active processes using the “Hard faults/sec” column in Resource Monitor. Determine which processes are most responsible for hard errors and RAM-related performance issues. These processes may consume excessive memory or fail to release it when they are no longer used.

Step 3: Implement Solutions

To remedy performance issues associated with RAM, consider the following steps:

– Distribute users across more session host VMs to effectively distribute RAM burden.

– Close or restart RAM-intensive processes if they are unnecessary or can be terminated securely.

– Instruct users to end their desktop sessions after each day

 Alternatively, implement policies that automatically log off users after a period of inactivity.

– Schedule VM restarts or use autoscaling to eliminate memory leakage and refresh the system.

 

Disk-Related Performance Limitations:

Disk performance is crucial for AVD environments to operate efficiently, and slow disk response and a lengthy disk queue can negatively impact the user experience. Here is a comprehensive method for diagnosing and resolving disk-related performance limitations:

 

Problem: Slow Disk Response and High Disk Queue Length

The local disk on session host VMs and FSLogix profile containers is essential for optimal system performance. Slow disk response and a lengthy disk queue can impede application launches and negatively influence the user experience overall. Follow these procedures to fix this problem:

 

Step 1: Assess the Local disk and FSLogix profile container.

Utilise Resource Monitor to monitor disc performance, focusing on the Disk Queue Length per volume. A consistently high queue length indicates that the operating system is spending considerable time waiting for disk responses, which degrades performance. Evaluate the local disk performance on session host virtual machines and FSLogix profile containers.

Step 2: Identify processes that generate disk activity.

Expand Resource Monitor’s Disk Activity section to identify processes causing excessive disk activity. This analysis assists in identifying the sources of excessive disk activity that contribute to a lengthy disk queue.

Step 3: Optimise Drive Performance

Consider the measures below to optimise disk performance and resolve disk-related issues:

– Ensure that file server virtual machines (VMs) hosting FSLogix profiles have adequate throughput and IOPS capabilities.

– Prefer Standard SSD (E-type) disks over Standard HDD storage at a minimum.

– Use Ephemeral OS discs for session host VMs to optimise storage performance.

– Use Premium SSD (P-type) disks on file servers that host FSLogix profiles, as these disks offer enhanced performance.

– Minimise disk I/O by roaming users’ indexed search cache data and cached Outlook data (OST file) within the FSLogix container.

 

Network-Related Performance Restrictions:

In AVD environments, network latency, packet loss, and bandwidth constraints can affect the user experience. Here is a comprehensive approach to diagnosing and resolving network-related performance limitations:

 

Problem: Issues with Latency, Packet Loss, and Bandwidth

It is essential to distinguish between bandwidth, latency, and packet loss to comprehend the specific network quality metrics influencing performance. To address network-related issues, follow these steps:

 

Step 1: Diagnose local connectivity issues

Perform speed experiments to evaluate the bandwidth of the local connection. You can measure latency and identify potential packet loss by pinging external hosts. Perform continuous ping tests while users are experiencing issues to identify instances of sporadic latency or packet loss.

Step 2: Analysis of Connection Segments

Identify the three network segments involved in AVD connectivity: the device of the user to the AVD Management Service, the AVD Management Service to the user’s Azure desktop VM, and the Azure desktop VMs to the public internet. Focusing on local connectivity, contacting Azure support for AVD Management Service issues, implementing corrective actions for VM-to-internet connectivity issues, troubleshooting and resolving issues specific to each segment.

Step 3: Determine the Required Bandwidth 

Determine the required bandwidth for virtual desktops based on the average consumption per active user and the potential peaks for multimedia applications. Size network connectivity appropriately to ensure a satisfactory user experience.

 

By following these steps, you can effectively troubleshoot and rectify performance limitations in Azure Virtual Desktop (AVD) environments. Taking care of CPU, RAM, disc, and network issues will optimise AVD performance, reduce user frustration, and ensure a successful desktop virtualisation deployment. With a thorough comprehension of these components and a methodical approach to problem resolution, you can create an AVD environment that meets the needs of your users.

Plex IT can offer invaluable assistance in optimising the AVD environment of an organisation. With their expertise and experience in Azure Virtual Desktop deployments, Plex IT can assist in identifying and resolving CPU, RAM, disc, and network performance issues. They can conduct exhaustive evaluations, implement best practices, and provide customised solutions to improve the overall performance and user experience within an AVD environment. By partnering a Managed IT Service Provider like Plex IT, businesses can utilise their expertise and direction to establish an AVD infrastructure that maximises productivity and meets their unique requirements. Contact us here.