Page tree
Skip to end of metadata
Go to start of metadata

The Clusters tab helps you identify the number of additional VMs that can be added to each cluster, the total and available resources in the cluster, and manages the resources of all the hosts it contains. It also enables you to view VMs that are currently powered on and running, their CPU and memory utilization, how many additional VMs can be added to each cluster depending on the spare cluster capacity, among other details.

Data displayed in the Clusters tab provides a complete picture of cluster memory usage and contention using Memory Metrics, and helps answer critical questions related to sub-systems, like:

  • Is vSphere memory management working right in my environment?
  • Is vSphere page sharing mechanism used effectively in my environment? How much memory is shared?
  • How much memory is over-committed?

This topic contains the following sections:

Viewing subsystem data in the Clusters tab

The Clusters tab is divided into the following sub-system views: 

ViewDescription
CapacityDisplays consolidated trend and saturation related metrics for the three subsystems: CPU, Memory and Storage. Also displays VM, Spare VM and Bottleneck related data.
CPUDisplays CPU related metrics for clusters.
MemoryDisplays memory related metrics, including Ballooning, Swap Rate and Granted memory for clusters.
StorageDisplays storage related data for vCenter clusters.

The following figure shows an example of the vSphere Clusters tab displaying Capacity view (default) metrics in the vSphere Cluster Used and Available Capacity view table.

vSphere Infrastructure view: Clusters tab

Viewing metrics in subsystem views

Different tables in the Clusters tab display detailed metrics and information for the following subsystems:

Viewing capacity metrics

Capacity view is default, and is displayed each time you access the Clusters tab.

Each row in the vSphere Cluster Used and Available Capacity view table represents a cluster and displays the following Capacity metrics.

By default, the metric tables in this view are filtered on all Domains. Use the Domain filter to select the required domain from the table.

By default, the Summarization Time Range for all metrics is set to Last 30 days. Use the Time Range controls to select an alternate range. The available options are:

  • Yesterday: Display data for the previous day, excluding today.
  • Last 7 days: Display data for the last 7 days, excluding today.
  • Last 30 days (Default): Display data for the last 30 days, excluding today

Capacity view: Description of columns

Column

Description

Statistical
Type 
 Metrics

Cluster

Name of the the vSphere cluster.

Last valueNAME
LocationLocation of the cluster.Last valueLOCATION_SITE
DatacenterName of the Datacenter associated with the cluster.Last valueRelationship DC_CONTAINS_VHC is used to get the NAME of the datacenter
vCenterName of the vCenter associated to the cluster.Last valueVIRTUALCENTER_NAME

CPU Utilization [%]

CPU utilization of the cluster computed by scaling the CPU MHz usage of the cluster by the effective CPU MHz available to the cluster.95th percentileCPU_UTIL 
CPU [MHz]Total CPU available/configured on the the cluster, measured in Megahertz.Last valueCPU_TOTAL_MHZ
CPU Pressure Index

A value between 0 and 100, calculated using CPU utilization and CPU ready time. This value must always be greater than or equal to CPU Utilization %. If this is much higher than CPU Utilization %, then it indicates that some VMs in the cluster are starved for CPU (high CPU ready time). See Virtual Machine view > Watchlist (VM) to identify them.

For more information, see Computation details for VMware metrics.

95th percentileIND_CPU_PRESSURE

CPU Daily Trend [MHz]

Daily growth rate of CPU utilization of the vSphere cluster, measured in Megahertz/day.

TrendComputed using CPU_UTIL_MHZ over time
CPU Days To SaturationEstimated number of days before the CPU capacity of the cluster breaches the configured threshold; if greater than 30 days, it will not be displayed.  NA IND_DAYS_TO_SATURATION[CPU] 
Memory [GB]Total physical memory configured for the cluster, measured in Gigabytes.Last valueTOTAL_REAL_MEM

Memory Utilization [%]

Percentage of total memory utilized by the vSphere cluster.

95th percentile MEM_UTIL
Composite Memory Utilization [%]

Consumed memory of a vSphere host or cluster is too high, while Active memory gives too low a picture of usage. The Composite memory consumption is a more realistic estimate of the amount of memory that is effectively in use in the host or cluster. It considers both Active and Consumed memory numbers reported by vSphere, and adjusts these using memory pressure due to contention, as measured by swapping, ballooning, and the total memory over-commitment in the host/cluster. Composite Memory Utilization [%] is a percentage calculated over the total memory of the host or cluster.

By default, spare VM calculation for a cluster in the vSphere infrastructure view uses memory utilization based on Consumed memory. The vSphere view has a parameter called "Spare VM calculation" in the Settings (vSphere) page by which you can use this Composite memory consumption instead for calculating spare VMs.

For more information, see Computation details for VMware metrics.

95th percentile IND_COMPOSITE_MEMORY_UTILIZATION
Datastore [GB]Total storage configured for the datastores in the cluster, measured in Gigabytes.Last value TOTAL_DSTORE_SIZE
Host CountThe total number of hosts configured in the cluster.Last value HOST_NUM
VMs On Daily TrendDaily growth rate of powered-on VMs.TrendComputed using VM_POWERON_NUM over time
Total Spec RatingThe total SPEC benchmark performance rating of the cluster, based on the CPU model and number of CPU cores.Last value BYBENCHMARK_VALUE[SPECINTRATE2006]
Unused Spec RatingUnused SPEC rating of the cluster, based on CPU utilization and the total SPEC rating.95th percentileComputed using CPU_UTILMHZ, CPU_TOTAL_MHZ and BYBENCHMARK_VALUE[SPECINTRATE2006]

Memory Daily Trend [GB]

Daily growth rate of memory utilization of the vSphere cluster per the current trend, measured in Gigabytes/day.

TrendComputed using MEM_CONSUMED over time

Memory Days To Saturation

Estimated number of days before the Memorycapacity of the cluster breaches the configured threshold; if greater than 30 days, it will not be displayed.  NAIND_DAYS_TO_SATURATION[MEMORY] 

Datastore Space Utilization [%]

Percentage of total datastore space utilized by the vSphere cluster.

95th percentileTOTAL_DSTORE_UTIL

Datastore Daily Trend [GB]

Daily growth rate of datastore utilization based on robust linear regression.TrendComputed using TOTAL_DSTORE_USED over time

Datastore Days To Saturation

Number of days before the datastore saturates.N/AIND_DAYS_TO_SATURATION[STORAGE  

Powered On VMs

Total number of VMs currently powered on in the vSphere cluster.

Average

GM_ON_NUM

Spare VMs

Estimated number of additional VMs that can be placed in the cluster.  For details, see Spare VM Calculation.

Computed metric:

  • 95th percentile values are used for CPU Used at cluster and VM level
  • Last value is used for cluster total capacity
  • Only currently powered ON VMs are considered
IND_SPARE_VM[GLOBAL]

Spare VM basis

Specifies the basis for calculating the number of additional VMs that can be placed in the cluster: TYPICAL or CUSTOM. The calculation also factors in Datastore Utilization (Storage) to calculate Spare VMs. For more information, see Specifying custom VM templates to calculate Spare VM basis.

NA TYPICAL or CUSTOM

Limiting Resource

The resource that limits the number of additional VMs that can be placed in the cluster. Limiting resource may not necessarily exceed threshold, it is the least available of the three resources – CPU, MEMORY and STORAGE.Computed using 95th percentile valuesResource with lowest Spare VMs among  IND_SPARE_VM[CPU], IND_SPARE_VM[MEMORY], IND_SPARE_VM[STORAGE]

For more information on VMware metics, see Metrics collected by VMware ETLs.

For more information on the Days to Saturation and other indicator parameters used for metric computations in this view, see Indicators.

Specifying custom VM templates to calculate Spare VM basis

To specify or edit typical or custom VM template settings to calculate Spare VM basis for an average VM, perform the following task:

  1. Under the Spare VM basis column in the table, click TYPICAL or CUSTOM.
  2. Under Specify custom VM template, select any one of the following options:
    1. Use average VM for cluster <cluster name> based on the past 30 days: Use the default (typical) VM template to calculate Spare VM basis for the selected cluster.
    2. Specify custom VM data for cluster <cluster name>: Enter values for the following attributes to calculate Spare VM basis for the selected cluster –

      AttributeDescription
      Custom VM name(Auto-populated) Spare VM Template
      Average CPU utilization [%]Average CPU utilization percentage value for the average VM.
      CPU core clock [MHz]Total CPU Megahertz value to use for the average VM.
      Number of CPUsNumber of CPUs to assign to the average VM.
      Total memory [GB]Total memory to assign to the average VM, measured in in Gigabytes.
      Average memory utilization [%]Average memory utilization percentage value to use for the average VM.
      Average datastore utilization [GB]Average datastore utilization to use for the average VM, measured in in Gigabytes.
  3. Click Save.
    • Depending on the values you have entered, your custom VM template might take some time to save.
    • An information banner under Specify custom VM template that says Custom VM saved successfully denotes a successful save.
    • Click Delete to delete the current custom VM template and return to the table.

Spare VM Calculation

Spare VMs are computed using the resource consumption of an average VM in the cluster. Average VM can be a typical VM in the cluster or a custom VM. The following table lists formulas for Spare VM calculation.

UtilizationFormula
Number of VMs that will fit in the cluster on the basis of the CPU needs of an average VM

SpareVM[CPU] = ((total_cpu_mhz * threshold[cpu_util]) - cpu_utilmhz) / cpu_of_average_vm

Expression used in the above formulaDescription

total_cpu_mhz

Total CPU capacity of the cluster (latest sample)

cpu_util

CPU utilization of the cluster (95th percentile)

cpu_utilmhz

Total CPU consumed by the cluster in MHz

cpu_of_average_vm

cpu_of_typical_vm or cpu_of_custom_vm

cpu_of_typical_vm

(cpu_utilmhz/gm_on_num)
cpu_of_custom_vmUser-specified value
Number of VMs that will fit in the cluster on the basis of the memory needs of an average VM

SpareVM[MEMORY] = ((total_real_mem * threshold[mem_util]) – memory_consumed) / memory_of_average_vm

Expression used in the above formulaDescription

memory_of_average_vm

memory_of_typical_vm or memory_of_custom_vm

total_real_mem

Total physical memory configured for the cluster (latest sample)
mem_utilPercentage of total memory utilized by the vSphere cluster (95th percentile)

memory_consumed

Memory consumed by the cluster, measured in Gigabytes

memory_of_typical_vm

(memory_consumed/gm_on_num)

memory_of_custom_vm

User-specified value
Number of VMs that will fit in the cluster on the basis of the storage needs of an average VM

SpareVM[STORAGE]=((total_dstore_size * threshold[total_dstore_util] - total_dstore_used /storage_of_average_vm

Expression used in the above formulaDescription

total_dstore_size

Total size of the datastore (latest sample)
total_dstore_utilPercentage of total datastore space utilized by the cluster

total_dstore_used

Total datastore space used by the cluster

Storage_of_average_vm

total_dstore_used/gm_on_num
Spare VM

SpareVM = MIN (SpareVM[CPU], SpareVM[MEMORY], SpareVM[STORAGE])

Limiting Resource

CPU, MEMORY, or STORAGE based on:

MIN (SpareVM[CPU], SpareVM[MEMORY], SpareVM[STORAGE])

Viewing CPU metrics

Each row in the vSphere Clusters CPU view table represents a cluster and its corresponding vCenter, and displays the following CPU metrics.

By default, the metric tables in this view are filtered on all Domains. Use the Domain filter to select the required domain from the table.

By default, the Summarization Time Range for all metrics is set to Last 30 days. Use the Time Range controls to select an alternate range. The available options are:

  • Yesterday: Display data for the previous day, excluding today.
  • Last 7 days: Display data for the last 7 days, excluding today.
  • Last 30 days (Default): Display data for the last 30 days, excluding today

CPU view: Description of columns

Column

Description

Statistical
Type
Metrics 

Cluster

Name of the vSphere cluster

Last valueNAME 

vCenter

Name of the vCenter on which the vSphere cluster is configured.

Last valueVIRTUALCENTER_NAME

CPU Utilization [%]

CPU utilization of the cluster computed by scaling the CPU MHz usage of the cluster by the effective CPU MHz available to the cluster.

95th percentileCPU_UTIL 
CPU Core clock MHzTotal processing capacity available to the vSphere cluster, measured in Megahertz.Last value
CPU_TOTAL_MHZ 

#Cores

Total number of physical CPU cores available to the vSphere cluster.

Last value
CPU_NUM
Unused CoresNumber of unused CPU cores.Computed metric- Based on 95th Percentile of CPU usage

Computed using CPU_UTIL and CPU_NUM

vCPUs

Total number of virtual CPU cores available to the vSphere cluster.

Last valueVCPU_NUM

#vCPU ON

Total number vCPUs that are currently powered on in the vSphere cluster.

Last valueVCPU_ON_NUM

#vCPU/host

Total number of vCPUs available to each host in the vSphere cluster.

Last valueComputed using VCPU_NUM and HOST_NUM

vCPU density

Total vCPU density, that is, total number of vCPUs per physical core, in the vSphere cluster.

Computed metric - uses average value for VCPU_NUM and CPU_NUMcomputed using VCPU_NUM and CPU_NUM
VMs with 1 vCPUNumber of VMs in the vSphere cluster that have 1 vCPU.Last valueComputed using VM.CPU_NUM
VMs with 2 vCPUsNumber of VMs in the vSphere cluster that have 2 vCPUs.Last valueComputed using VM.CPU_NUM
VMs with >2 vCPUsNumber of VMs in the vSphere cluster with more than 2 vCPUs.Last valueComputed using VM.CPU_NUM

For more information on VMware metics, see Metrics collected by VMware ETLs.

The following image shows an example of the vSphere Clusters CPU view table.

vSphere Clusters CPU view table

Viewing memory metrics

Each row in the vSphere Clusters Memory view table represents a cluster and its corresponding vCenter, and displays the following memory metrics.

By default, the metric tables in this view are filtered on all Domains. Use the Domain filter to select the required domain from the table.

By default, the Summarization Time Range for all metrics is set to Last 30 days. Use the Time Range controls to select an alternate range. The available options are:

  • Yesterday: Display data for the previous day, excluding today.
  • Last 7 days: Display data for the last 7 days, excluding today.
  • Last 30 days (Default): Display data for the last 30 days, excluding today

Memory view: Description of columns

Column

Description

Statistical
Type
 Metrics

Cluster

Name of the vSphere cluster

Last valueNAME

vCenter

Name of the vCenter on which the vSphere cluster is configured.

Last valueVIRTUALCENTER_NAME 
DatacenterName of the datacenter on which the vSphere cluster resides.Last valueRelationship DC_CONTAINS_VHC is used to get the NAME of the datacenter

Total Memory [GB]

Total amount of memory available to the vSphere cluster, measured in in Gigabytes.

Last valueTOTAL_REAL_MEM
Memory Over-commitment [%]

Sum of memory configured for all the VMs in the cluster divided by the total memory of the cluster expressed as a percentage. If this number is over 100%, then the memory is over-committed. Numbers over 100% are not necessarily an indication of a problem, it indicates efficient use of virtualized resources. However, extremely high memory overcommitment could result is performance problems. The amount of over-commitment can vary in each environment based on the nature of applications and staggering of applications running on the VMs. The default low and high thresholds are 100% and 120%. You can set appropriate thresholds for your environment from the Settings page.

Computed metric - based on Last valueComputed using SUM(VM.TOTAL_REAL_MEM) and Cluster.TOTAL_REAL_MEM

Memory Utilization [%]

Memory usage scaled by total memory of the cluster. Usage is based on consumed memory.

95th percentile MEM_UTIL
Composite Memory Utilization [%]

Consumed memory of a vSphere host or cluster is too high, while Active memory gives too low a picture of usage. The Composite memory consumption is a more realistic estimate of the amount of memory that is effectively in use in the host or cluster. It considers both Active and Consumed memory numbers reported by vSphere, and adjusts these using memory pressure due to contention, as measured by swapping, ballooning, and the total memory over-commitment in the host/cluster. Composite Memory Utilization [%] is a percentage calculated over the total memory of the host or cluster.

By default, spare VM calculation for a cluster in the vSphere infrastructure view uses memory utilization based on Consumed memory. The vSphere view has a parameter called "Spare VM calculation" in the Settings page by which you can use this Composite memory consumption instead for calculating spare VMs.

For more information, see Computation details for VMware metrics.

95th percentileIND_COMPOSITE_MEMORY_UTILIZATION 

Active Memory [%]

Sum of active memory for all powered-on virtual machines. Active memory of the virtual machine is the amount of memory that is actively used, as estimated by VMkernel based on recently touched memory pages.

95th percentileComputed using ACTIVE_MEMORY and TOTAL_REAL_MEM 

Balloon Memory [%]

The sum of balloon memory of all powered-on virtual machines. Balloon memory of a virtual machine is the amount of physical memory reclaimed through ballooning.

95th percentileComputed using BALLOONING_MEMORY and TOTAL_REAL_MEM 

Granted Memory [%]

The percentage of machine memory that the VMkernel has given to the vSphere cluster based on its memory demand.

95th percentileComputed using GRANTED_MEMORY and TOTAL_REAL_MEM 

Shared Memory [%]

Sum of all shared metrics for all powered-on virtual machines. For a VM, it is the amount of “physical” memory shared with other virtual machines through the VMkernel’s transparent page-sharing mechanism.

95th percentileComputed using SHARED_MEMORY and TOTAL_REAL_MEM  
Overhead memory [%]Total of all overhead metrics for powered-on virtual machines. For a VM, it is the amount of machine memory used by the VMkernel to run the virtual machine.95th percentileComputed using OVERHEAD_MEMORY and TOTAL_REAL_MEM  
Memory Swap Rate [MB/s]Sum of the rate at which memory is being swapped out from active memory to disk and swapped in from disk into active memory. It is measured in megabytes/second.95th percentileMEM_SWAP_IN_BYTE_RATE, MEM_SWAP_OUT_BYTE_RATE

For more information on VMware metics, see Metrics collected by VMware ETLs.

The following image shows an example of the vSphere Clusters Memory view table.

vSphere Clusters Memory view table

Viewing storage metrics

Each row in the vSphere Clusters Storage view table represents a cluster and its corresponding vCenter, and displays the following storage metrics.

By default, the metric tables in this view are filtered on all Domains. Use the Domain filter to select the required domain from the table.

By default, the Summarization Time Range for all metrics is set to Last 30 days. Use the Time Range controls to select an alternate range. The available options are:

  • Yesterday: Display data for the previous day, excluding today.
  • Last 7 days: Display data for the last 7 days, excluding today.
  • Last 30 days (Default): Display data for the last 30 days, excluding today

Storage view: Description of columns

Column

Description

Statistical
Type
Metrics 

Cluster

Name of the vSphere cluster.

Last valueNAME 

vCenter

Name of the vCenter on which the vSphere cluster is configured.

Last valueVIRTUALCENTER_NAME 

Datastore Count

Total number of datastores present in the vSphere cluster.

MAXComputed based on the count of datastore entities associated with the cluster

Capacity [TB]

Total storage capacity of the datastore, measured in terabytes.

Last valueTOTAL_DSTORE_SIZE

Utilization [%]

Usage of all the datastores in the cluster scaled by the storage capacity of the cluster, expressed in percentage.

95th percentileTOTAL_DSTORE_UTIL

Used [TB]

Total storage space used by the datastore, measured in terabytes.

95th percentileTOTAL_DSTORE_USED1

Free [TB]

Free storage capacity of the cluster. This value is computed using the formula 'Capacity' – 'Used'

95th percentileTOTAL_DSTORE_FREE1

Provisioned [TB]

Sum of the storage provisioned for all the datastores in the cluster. Just as memory storage can be over-provisioned, a small amount of over-provisioning indicates efficient use of resources. Excessive over-provisioning can cause the cluster to run out of storage resources.

Last valueTOTAL_DSTORE_PROVISIONED1

1 Metrics are available only for clusters that have shared datastores in them.

For more information on VMware metics, see Metrics collected by VMware ETLs.

The following image shows an example of the vSphere Clusters Storage view table. 

vSphere Clusters Storage view table

Viewing vSphere cluster details

To view the details page for a particular vSphere cluster, click its (hyper-linked) name under the Cluster column in any of the view tables.

The details page for each cluster displays the following information:

  • Cluster details: General cluster details, such as name, type, description, and tags

  • Configuration data: Cluster configuration data table. The configuration data displayed under the Resource and Value columns in this table varies for each sub-system.

  • Related information: Links to detailed pages for Thresholds, Host details, VM details and VM Analysis. For details, see Viewing related information.

  • Recommendations: All available Recommendations for this cluster. The Recommendations table for the selected cluster displays the following information:

    ColumnDescription
    Icon

    Severity/Efficiency icon

    TypeDisplays all problem types for this system, for example, Forecasted saturation, Overallocated VMs.
    ResourcesDisplays all affected resources (sub-systems), for example, Memory, CPU.
    DescriptionA brief description of all problem types. For example, Few resources are saturated, VM is overallocated.
    Recommended actionsActions to take to resolve the problem.
    View recommendation detailsView recommendation details for the selected problem type in a pop-up box.

    For detailed information, see Recommendations (vSphere).

  • Analysis charts: Core metrics for the selected entity are displayed as analysis charts, depending on available data. At the bottom of each chart is a link that takes you to the Data Explorer view for the specific resource type, and displays charts and information for a larger set of metrics.

    To return to this page, click the <<Go to Detail Page button located towards the top-right side in the Data Explorer tab.

    The data displayed in Data Explorer view charts for any resource part of this view category, is based on the time filter that is applied to the primary view in this tab. You can select Advanced time periods (other than default/basic) from the Data Explorer view for the selected resource, to view charts over a different time resolution.


    For more information, see Data Explorer view for vSphere Cluster.

The following figure shows an example of the details page for a cluster accessed from Storage view.

Cluster details page: An example of detailed information for a particular cluster

Viewing related information

The Related Information section above the Recommendations table in the cluster details page contains the following links:

LinkDescription
ThresholdsThe thresholds page allows you to view and manage threshold settings for the selected vSphere cluster. See Thresholds for detailed information.
Host detailsClicking on Host details takes you to the Host View tab, that displays information about all configured hosts in the selected vSphere cluster. See Hosts for more information.
VM detailsClicking on VM details takes you to the All VMs tab in the Virtual Machine View category, that displays information about all the VMs in the selected vSphere cluster. See All VMs for more information.
VM Analysis

Clicking on VM Analysis displays color-coded graphs for all applicable sub-system: CPU, Storage, Memory and Datastore utilization.

Using data filter and other controls in tables

You can export and filter the information that is displayed in tables to suit your need. For instance, you can show or hide columns, filter tables on particular domains, and so on.

For more information, see Exporting and filtering tables.

4 Comments

  1.  

  2.  

    1.  

  3.