CPU ready chart

CPU Ready is the measurement of time that a VM is ready to use CPU, but is unable to schedule physical CPU time because host CPU resources are busy.

Turbonomic collects CPU Ready data from hosts and VMs that are discovered by vCenter targets. It then calculates CPU Ready capacity and utilization to make accurate VM move recommendations.

When you set the scope to one or several VMs or hosts, the CPU Ready chart displays CPU Ready data.

CPU Ready chart

Host CPU ready capacity

Turbonomic calculates host CPU Ready capacity by using following formula:

Host Logical Processors * 20 = Host CPU Ready Capacity
The following values are represented in the formula:
  • Host Logical Processors is the number of logical CPU cores on a host.

  • 20 is the standard Ready Queue interval (in seconds) at which the hypervisor measures metrics.

For example, if a host has 10 cores and applies the default host policy, Turbonomic calculates capacity according to the following formula:

10 * 20 = 200

Host CPU ready utilization

CPU Ready utilization is the measurement of capacity that is in use. Turbonomic calculates host CPU Ready utilization by using following formula:
Host CPU Ready Average / Host CPU Ready Capacity = Host CPU Ready Utilization
The following values are represented in the formula:
  • Host CPU Ready Average is the average of the 20-second CPU Ready summation values that are collected from vCenter every 10 minutes. Each value is expressed in number of milliseconds.

  • Host CPU Ready Capacity is the capacity that is calculated by Turbonomic. For more information, see the previous section.

For example, if raw utilization on a host is 40 and CPU Ready capacity is 100, Turbonomic uses the following calculation for host CPU Ready utilization.

First, to get the Host CPU Ready Average value, Turbonomic collects 20-second CPU Ready summation values from vCenter during the poll period, for example:

AVERAGE(500, 100, 1000, 1500, 1500, 1500, 500, 100, 1000, 1500, 1500, 1500, 500, 100,
1000, 1500, 1500, 1500, 500, 100, 1000, 1500, 1500, 1500, 500, 100, 1000, 1500, 1500,
1500) = 920 ms = 0.92 s
Turbonomic then divides the Host CPU Ready Average value by Host CPU Ready Capacity, which in this example is 100, to get a host CPU Ready utilization value of 0.92, or about 9%
0.92/100 = 0.0092 (0.9%)

To view the hosts with the highest CPU Ready, add the Top Utilized chart to your dashboard. When you configure the chart, select Hosts as the entity type, and CPUReady as the commodity.

To view the host clusters with the highest CPU Ready, add the Top Utilized chart to your dashboard. When you configure the chart, select Host Clusters as the entity type and CPUReady as the commodity.

Top Utilized Host Clusters chart

This chart shows any pending action to move VMs out of a host cluster due to CPU Ready congestion.

VM CPU ready capacity

Turbonomic calculates VM CPU Ready capacity by using the following formula:

VM Logical Processors * 20 = VM CPU Ready Capacity
The following values are represented in the formula:
  • VM Logical Processors is the number of vCPUs on a VM.

  • 20 is the standard Ready Queue interval in seconds at which the hypervisor measures metrics.

For example, for a VM with 2 vCPUs, Turbonomic uses the following calculation for VM CPU Ready capacity:

2 * 20 = 40

VM CPU ready utilization

CPU Ready utilization is the measurement of capacity that is in use. Turbonomic calculates VM CPU Ready utilization by using following formula:

Raw Utilization / VM CPU Ready Capacity = VM CPU Ready Utilization
The following values are represented in the formula:
  • Raw Utilization is the utilization value (in seconds) that is collected from vCenter.

  • VM CPU Ready Capacity is the capacity that is calculated by Turbonomic. For more information, see the previous section.

For example, if raw utilization on a VM is 40 and CPU Ready capacity is 100, Turbonomic uses the following calculation for VM CPU Ready utilization:

40/40 = 1 (100%) 

To view the most utilized VMs, add the Top Utilized chart to your dashboard. When you configure the chart, select virtual machines as the entity type, and CPUReady as the commodity.

Top utilized VMs chart

This chart shows any pending action to move a VM to a different host due to CPU Ready congestion.

Effect on VM move sctions

Turbonomic considers host CPU Ready utilization when it makes placement decisions for VMs. Ready Queue Utilization is a host policy setting for the percentage of utilization that Turbonomic considers as full utilization. For example, if utilization reaches 50%, Turbonomic considers Ready Queue to be fully utilized and the market might generate a move action to remedy the high Ready Queue utilization condition. The default value for this policy setting is 50%. This percentage is suitable for most environments. For environments where applications are sensitive to latency, you can reduce the percentage so that Turbonomic is more sensitive to CPU Ready when it places VMs. However, CPU Ready utilization is only one among many factors that go into move decisions. In some environments, other factors might offset CPU Ready utilization concerns.

Logical Processors are considered when Turbonomic places VMs on hosts. For example: If a VM has 64 vCPUs, it must run on a host with at least 64 logical cores. Turbonomic does not attempt to move a VM to a host with fewer Logical Processors. For VMs, Logical Processors Capacity and Used values equal the number of vCPUs that are configured on the VM. For hosts, Logical Processor Capacity equals the number of host logical cores and the Used value equals the sum of all VM Logical Processors (vCPUs). In this way, the Host Logical Processor Utilization represents the ratio of VM vCPUs to the host logical cores. Overprovisioned hosts are known to contribute to increased risk of CPU Ready latency. Be sure to take any resize down actions on VMs where CPU Ready is observed to reduce CPU overprovision.

Note: In VMware environments, the best practice is to keep CPU Ready values as low as possible. A CPU Ready value of 3% indicates a potential performance risk for most applications. If the CPU Ready value is 5% or greater, expect a significant performance impact.

CPU Ready Dashboard: identifying vPU ready risks

A default dashboard is available for reviewing CPU Ready in your environment. This dashboard shows the Host Clusters, Hosts, and VMs with the highest CPU Ready values. Reviewing the charts in this dashboard can help you tune your CPU Ready settings for the specific workloads that are running in your environment. To view this dashboard, select DASHBOARD from the main navigation menu and click On-Prem CPU Ready Dashboard.