Monitor Performance of VMs, Hosts, Clusters

Using Commander, you can analyze how a VM, host, or cluster is performing and what issues may be affecting its performance. Performance metrics are retrieved from the cloud provider. By default, Commander gathers performance data from vCenter and SCVMM and can display metrics on the Performance tab for VMs, hosts, and clusters. Additional configuration may be required for public cloud accounts.

Commander monitors CPU, memory, disk, and network usage over the last seven days. The average number is calculated by averaging the resource usage over the last seven days. Commander also collects information about the maximum, or peak resource usage over the last seven days.

The last seven days of performance information is retrieved when a cloud account is added to Commander. SCVMM is the exception to this rule. For SCVMM, because historical performance data can't be retrieved from the cloud account, Commander waits to obtain seven days of performance data from SCVMM before generating a performance summary.

VM performance metrics are updated automatically by Commander every night. You can also manually update performance data for individual VMs.

GCP prerequisite: To allow Commander to obtain performance data for GCP VMs, the service account used to add GCP cloud account must have at least the Monitoring Viewer role. To learn more, see Grant permissions to the Commander service account.

Important: Memory usage monitoring requires additional configuration for public cloud instances. See:

VM performance metrics

The following performance metrics are available in the Weekly Performance Summary for a VM.

Memory usage monitoring requires additional configuration for public cloud instances.

Metric

Description

Available for

CPU Usage (%)

Amount of actively used virtual CPU, as a percentage of total available CPU.

AWS, Azure, GCP, SCVMM, vCenter

CPU Usage (MHz)

Amount of actively used virtual CPU. This is the host's view of the CPU usage, not the guest operating system view.

vCenter

CPU Ready (%)

Percentage of time that the VM was ready, but couldn't be scheduled to run on the physical CPU. CPU ready time is dependent on the number of VMs on the host and their CPU loads.

vCenter

Memory Usage (%)

Percentage of allocated memory.

Note for SCVMM: Memory performance information is unavailable for VMs configured with static memory, because SCVMM always represents the memory usage as 100%.

AWS, Azure, GCP, SCVMM, vCenter

Memory Consumed (MB)

Amount of physical memory consumed by the VM.

  • For private cloud, consumed memory doesn't include overhead memory. It includes shared memory and memory that might be reserved, but not actually used.
  • For public cloud Linux instances, the consumed memory is the "user" memory and doesn't include other categories such as buffer and cached memory.
  • For GCP Windows instances, consumed memory is calculated using (Total Memory * Usage Percentage).
  • For SCVMM, memory performance information is unavailable for VMs configured with static memory, because SCVMM always represents the memory usage as 100%.

AWS, Azure, GCP, SCVMM, vCenter

Memory Overhead (MB)

Amount of machine memory used by the VM kernel to run the VM.

vCenter

Memory Ballooning (MB)

Amount of guest physical memory that's currently reclaimed from the VM through ballooning. This is the amount of guest physical memory that has been allocated and pinned by the balloon driver.

vCenter

Disk Usage (MB/s)

Aggregated disk I/O rate.

  • For public cloud, disk metrics are an aggregate of metrics for all disks on the VM. However, Commander doesn't display Disk Usage for AWS instance store-backed volumes, and for Azure, disk metrics are retrieved only for OS disks.
  • For vCenter and SCVMM hosts, this metric includes the rates for all VMs running on the host during the collection interval. Not available for NFS.

AWS, Azure, GCP, SCVMM, vCenter

Disk Max Latency (ms)

Highest latency value for any operation executed by the VM. Latency measures the time taken to process a SCSI command issued by the guest OS to the VM.

vCenter

Disk Swap In (MB)

Rate at which memory is swapped from disk into active memory during the interval.

vCenter

Disk Swap Out (MB)

Rate at which memory is swapped from active memory to disk during the interval. This counter is generally more useful than the swap-in counter to determine whether the VM is running slowly due to swapping, especially when looking at real-time statistics.

vCenter

Disk IOPS

Sum of disk read and write IOPS.

For Azure, disk metrics are retrieved only for OS disks. This value is a sum of Disk Read Operations/Sec and Disk Write Operations/Sec.

For GCP, this value is a sum of compute.googleapis.com/instance/disk/read_ops_count and compute.googleapis.com/instance/disk/write_ops_count.

Azure, GCP

Network Usage (MB/s)

Sum of data transmitted and received across all virtual NIC instances connected to the VM.

AWS, Azure, GCP, SCVMM, vCenter

View weekly VM performance summaries

Use a VM's Performance Summary tab to view a summary of the last seven days of VM performance data. VM performance metrics are updated automatically by Commander every night.

Access:

Views > Inventory > Infrastructure or Applications

Available to:

View performance metrics: All Access Rights Levels

Update performance metrics: All Operator Levels and Higher

  1. From the inventory tree, select a VM.
  2. Select the Performance Summary tab.
  3. To update the performance summary for a VM so that it includes the most recent performance data, select Update Performance then OK.

Analyze vCenter performance

The following vCenter performance charts are only available when the Time Range is set to 1 Hour:

  • Disk Rate (KBps)
  • Disk Requests (Number)
  • Network Rate (Mbps)
  • Network Requests (Number)

The following vCenter performance charts are only available when the Time Range is greater than 1 Hour:

  • Disk (KBps)
  • Network (MBps)
  • Space (GB) — only available for VMs

Access:

Views > Inventory > Infrastructure or Applications

Available to:

View performance metrics: All Access Rights Levels

Update performance metrics: All Operator Levels and Higher

  1. From the inventory tree, select a vCenter VM, host, or cluster.
  2. Select the Performance Charts tab.

    The performance charts are loaded the first time you click the Performance Charts tab for the selected VM, host, or cluster, and default to the time range of one hour.

  3. Optional: To see the performance over different periods of time, set the Time Range as required.

    All performance charts are updated to display the same time range.