Virtualization Manager Reports > VM Performance Summary
  
Version 10.2.01P10
VM Performance Summary
To access this report, in the Navigation pane select:
When a VM exhibits performance issues, it could stem from either the server, the actual storage, or other network factors. Once server bottlenecks are addressed, the next step is to determine if there’s an issue with the LUN and disks. While vCenter provides performance statistics, this information is not always readily available to Storage Administrators.
The VM Performance Summary enables a Storage Administrator to have an at-a-glance look at I/O performance statistics: CPU, Memory, and Disk I/O. In addition, Disk I/O is broken out into reads and writes, with latency statistics. In addition, this report provides the necessary CPU Ready States and CPU Wait States to determine if an ESX host can handle another VM Guest.
Disk Read and Write Latency is available only for ESX Hosts (not VMs). Therefore, latency performance charts are available only for partitions of Type=Host.
In the vCenter VM ESX Server settings, the statistics level must be set to Level 4 in order for balloon statistics to appear in the VM Performance Summary report. When Level 4 is configured, it will also capture the other statistics, as listed in the following table, Performance Counters and Statistics Level Settings. See vCenter Statistics Configuration.
In the vCenter VM ESX Server settings, the statistics level must be set to the maximum level based on the counter listed in the following table, Performance Counters and Statistics Level Settings. Note that level 1 is always captured, but other levels may need to be set for other performance counters. Configuring a level lower than the required level for a counter will result in a failure to retrieve the relevant data from VSphere and will cause inaccurate values in reports. See vCenter Statistics Configuration.
Table 1 Performance Counters and Statistics Level Settings
Performance Counters Collected
Statistics Level
Memory (%)
2
Memory Usage (%)
1
Avg CPU Usage (%)
1
Avg CPU Usage (KB)
1
Disk Avg Usage (KB/sec)
1
Disk Reads (KB/sec)
2
Disk Writes (KB/sec)
2
Disk # of Reads
3
Disk # of Writes
3
Disk Total Latency (ms)
2
Disk Read Latency (ms)
2
Disk Write Latency (ms)
2
Disk Queue Latency (ms)
2
Disk Queue Read Latency (ms)
2
Disk Queue Write Latency (ms)
2
Maximum Queue Depth
1
vCenter Statistics Configuration
The following example illustrates how statistics levels can be set in vCenter.
Focusing on Specific I/O Statistics
To isolate a particular set of statistics—for example, Disk I/O—use the Advanced options in the Report Designer and click the relevant checkbox.
All values represent the average for the interval selected in the “Group By” option.
Name
The name of the VM. Click on this link to view VM Server Details, if it is a host, or VM Detail, if it is a VM.
Type
Type of partition: host server or VM; Disk Read and Write Latency is available only for ESX Hosts (not VMs). Therefore, latency performance charts are available only for partitions of Type=Host.
CPU Usage
CPU %
The VM’s CPU usage percentage.
CPU Max %
Maximum CPU usage percentage for the report’s Group By interval.
Avg Wait %
Calculated for VM guests based on the total milliseconds for each data capture interval included in the reporting interval, divided by the number of milliseconds in the reporting interval; used to determine how long VM guests wait to get CPU.
Max Wait %
Calculated for VM guests based on the total milliseconds for each data capture interval included in the reporting interval, divided by the number of milliseconds in the reporting interval; used to determine how long VM guests wait to get CPU.
Avg Ready %
Calculated for VM guests based on the total milliseconds for each data capture interval included in the reporting interval, divided by the number of milliseconds in the reporting interval; used to determine how long VM guests wait to get CPU.
Max Ready %
Calculated for VM guests based on the total milliseconds for each data capture interval included in the reporting interval, divided by the number of milliseconds in the reporting interval; used to determine how long VM guests wait to get CPU.
CPU MHz
Average processor speed for the report’s Group By interval.
CPU Max MHz
Maximum processor speed for the report’s Group By interval.
Balloon Memory
Min (KB)
Minimum of all the data collection intervals within the report’s time period.
Avg (KB)
Average of all the data collection intervals within the report’s time period.
Max (KB)
Maximum of all the data collection intervals within the report’s time period.
Memory Usage
Memory %
Average memory usage percentage for the report’s Group By interval.
Memory Max %
Maximum memory percentage for the report’s Group By interval.
Memory MB
Average memory usage for the report’s Group By interval.
Memory Max MB
Maximum memory usage for the report’s Group By interval.
Disk I/O
Disk I/O Data Xfer (KB/sec)
Disk I/O average transfer rate for all disks for the VM.
Disk Read
Data Xfer (KB/sec)
Disk read transfer rate, average for the report’s Group By interval.
# Reads
Sum of the number of disk reads for the report’s Group By interval.
Disk Write
Data Xfer (KB/sec)
Disk write transfer rate for the report’s Group By interval.
# Writes
Sum of the number of disk writes for the report’s Group By interval.
Show Chart
Check the boxes for which you want to generate charts and click Show Chart to view the VM Performance Over Time graphs. Disk Read and Write Latency is available only for ESX Hosts (not VMs). Therefore, the latency performance charts are available only for partitions of Type=Host. See also Show Chart Parameter.
Show Chart Parameter
When selecting the Show Chart option in a VM performance report, a Group By parameter can be applied to the selection.
Note that the default scope selector time period is set to Last 7 Days, so if a Group By greater than Days is selected—that is, Weeks, Months, Quarters, or Years—a chart will not be created.