The most likely causes of the issue are A. Disk I/O limits and C. CPU oversubscription. Disk I/O limits are the maximum amount of input/output operations per second (IOPS) that a disk can handle. CPU oversubscription is the ratio of virtual CPUs to physical CPUs in a host. Both of these factors can affect the performance of a VDI environment, especially during peak hours when many users log in and launch applications.
Disk I/O limits can cause slow boot times, application lags, and cursor freezes for VDI users12. To avoid this issue, it is recommended to use flash storage or SSDs for VDI workloads, as they have much higher IOPS than traditional hard disk drives31. It is also important to monitor the disk performance and adjust the disk size and configuration as needed1.
CPU oversubscription can also cause performance degradation for VDI users, as it can lead to CPU contention and increased latency42. To avoid this issue, it is recommended to limit the CPU oversubscription ratio to a reasonable level, such as 4:1 or lower42. It is also important to monitor the CPU utilization and balance the load across hosts as needed4.
The other options are less likely to cause the issue. Affinity rules are used to specify which virtual machines should run on which hosts or which virtual machines should not run on the same host. They are not related to the performance of VDI workloads. RAM usage can affect the performance of VDI workloads, but it is usually not a major factor during peak hours, as most users do not consume a lot of memory when they log in or launch applications. Insufficient GPU resources can affect the performance of VDI workloads that require high graphics processing, such as video streaming or 3D rendering, but they are not relevant for most VDI users. License issues can affect the availability of VDI workloads, but they are not related to the performance of VDI workloads.
