Stressed Hosts

The report helps you detect resource demanding virtual machines that fail to operate because the hosts where these VMs reside are not capable to sustain the workload.

Description

The Stressed Hosts report analyzes historical performance, configured resource allocation and current workloads to provide recommendations for an optimized allocation of resources.

The report can be used in addition to the Virtual Machines. Right-sizing — VMs Oversized for Memory and CPU and Virtual Machines. Right-sizing — VMs Undersized for Memory and CPU reports to detect VMs not discovered by the right-sizing reports but still experiencing lack of available resources.

Parameters

Date/time Selection: defines a time period to analyze in the report.

Include week days: defines specific week days from the selected time period to analyze in the report.

Show details: defines whether the report should display results in the collapsed or expanded view. If you choose No, only general information will be displayed in the report. The report will not include charts in the output.

Threshold: Dispatch (%): defines CPU dispatch wait time threshold.

The CPU Dispatch Wait Time % is a percentage of time that a virtual machine was ready, but could not get scheduled to run on the physical CPU. If CPU dispatch wait time for a VM exceeds the specified threshold, the report will consider the VM to be resource demanding.

Scope: defines a list of Groups or Objects to include in the report (by default, the Veeam Hyper-V Hosts Group is selected). The Add Group option returns a Group and all objects that a group contains (with all performance counter instances). The Add Object option returns only the selected object.

Sections to show: defines whether the report will analyze RAM resource allocation, CPU resource allocation or both.

Top N Count: defines the maximum number of virtual machines on each stressed host to display in the report output.

Performance Modeling Based On: defines whether the report will use maximum (peak) or average CPU dispatch wait time historical values to assess VM performance.

Sample Usage

Report on Hyper-V hosts in your infrastructure that have performance issues and display top 5 resource demanding VMs running on the hosts, assuming that the maximum CPU dispatch wait time has to be lower than 15%. Use performance statistics for the previous quarter as a ground for the report analysis; omit statistics collected on Saturdays and Sundays.

Instructions

  1. Open the Stressed Hosts report.
  2. From the From list, select Previous quarter > First day. From the To list, select Previous quarter > Last day.
  3. In the Include week days field, clear check boxes next to Saturday and Sunday.
  4. From the Sections to Show list, select both vRAM and vCPU.
  5. From the Show Details list, select Yes (collapsed).
  6. In the Threshold: Dispatch (%) field, enter 15.
  7. In the Top N Count field, enter 5.
  8. From the Performance Modeling Based On list, select Average Usage.
  9. Click Run to view the report.

Report Output

In the Stressed Hosts: CPU and Stressed Hosts: Memory sections, the report will display a list of hosts that are running low on CPU and memory resources. For each stressed host, the report will show current host CPU configuration and summary information on the amount of total and available memory resources.

In the Stressed Hosts: Memory section, click the plus sign next to the name of a stressed host to see the list of stressed VMs, their current and recommended vRAM allocation, maximum memory pressure and maximum vRAM usage values. Click the plus sign next to the name of a stressed VM to analyze how the maximum memory pressure level and the amount of available memory resources have changed for the VM during the reporting interval.

In the Stressed Hosts: CPU section, click the plus sign next to the name of a stressed host to see the list of stressed VMs, their current vCPU allocation, maximum/average CPU dispatch wait time values and the percentage of physical processor time being consumed by these VMs. Click the plus sign next to the name of a stressed VM to analyze how the CPU dispatch wait time (maximum and average) values have changed for the VM during the reporting interval.

The RAM analysis and CPU analysis tables will show the total number of processed VMs, number of stressed VMs and VMs that cannot be analyzed.

Note

If a virtual machine was turned off at some time during the reporting interval, then the configuration information (amount of allocated memory) for this period may be unavailable.

All recommendations are based on the current configuration. If a virtual machine has no current configuration (is turned off or decommissioned), it will not be included in the report.

You can click the Stressed VMs link in the RAM analysis or CPU analysis table, or the name of a stressed VM in the Stressed Hosts: Memory or Stressed Hosts: CPU tables to drill down to performance details for these VMs.

You can click the VMs that cannot be analyzed link in the RAM analysis and CPU analysis tables to drill down to performance details for these VMs.