Policies enable the user to view and/or modify certain parameters that may place boundaries around the analytics engine or serve as inputs for certain features and capabilities.

Best Practices

The Best Practices thresholds – for CPU Ready Time, Memory Swapping and Latency – are utilized by SIOS iQ while it learns the underlying behaviors of newly added environments to identify issues according to the established industry standards captured in this policy. These default threshold values are read-only and apply globally across all defined environments.

  • CPU Ready Time (default = 300 ms)
  • Memory Swapping (default = 0 MB)
  • Latency (default = 25 ms)

Cost

Cost calculation values are used to calculate the infrastructure cost of your environments. The default Monthly Cost values are based on general industry trends in US Dollars. To specify a different Monthly Cost value or currency, select Custom and enter the desired cost unit. Cost calculation values are global and apply to all VMs in all environments. These values are used to calculate the potential monthly savings in the Efficiency dashboard.

The default Monthly Cost values internal to SIOS iQ are in monetary units and are based on general industry trends in US Dollars.

  • Monthly vCPU Cost (default = $40)
  • Monthly Memory Cost per GB (default = $10)
  • Monthly Storage Cost per GB (default = $.05)

Idle VMs

Idle VMs are listed on the Efficiency Dashboard. Idle VMs are identified using the values specified and the calculated average utilization of CPU, disk and network resources within the environment. Policy values are global and apply to all VMs in all environments. For more information, see Efficiency > Idle VMs.

These values are percentages. For CPU, it is a straight percentage of CPU usage. For disk and network usage, the values are percentages of the overall average of all the virtual machines in the environment.

  • Avg CPU Utilization (default = 10%)
  • Percent of Avg Disk Utilization (default = 10%)
  • Percent of Avg Network Utilization (default = 10%)

There are 3 parameters that are used to define Idle VMs (Avg CPU Utilization, Percent of Avg Disk Utilization, Percent of Avg Network Utilization). An Idle VM is identified when ALL THREE of its average utilizations are equal to or less than the average environment utilization calculated using the values specified. The user can set the time period used to calculate the utilization and identify Idle VMs directly within the Efficiency Dashboard.

Avg CPU Utilization

  1. Idle VMs are displayed on the Efficiency Dashboard when the average CPU utilization for a particular VM is equal to or less than the parameter set in the policy. (Ex. In this example the Avg CPU Utilization is set to 10%).
  2. Policy values are global and apply to all VMs in all environments.

Percent of Avg Disk Utilization

  1. Idle VMs are displayed on the Efficiency Dashboard when the average disk utilization for a particular VM is equal to or less than the idle value for disk utilization as calculated using the current policy setting. The idle value for Avg Disk Utilization is calculated by multiplying the percent value set in the policy and the Avg Disk Utilization of all of the VMs in the environment. (Ex: If there are a total of 3 VMs, with Disk Utilizations of 20 KBps, 50 KBps, and 30 KBps respectively, then the Avg Disk Utilization across all 3 VMs is 33.33KBps. With the policy value set to 10% the dashboard will only display Idle VMs that have a Disk Utilization of less than 3.33 KBps).
  2. Policy values are global and apply to all VMs in all environments.

Percent of Avg Network Utilization

  1. Idle VMs are displayed on the Efficiency Dashboard when the average network utilization for a particular VM is equal to or less than the value calculated using the current policy setting. Avg Network Utilization is calculated by multiplying the percent value set in the policy and the Avg Network Utilization of all of the VMs in the environments. (Ex: If there are a total of 3 VMs, with Network Utilizations of 10 KBps, 15 KBps, and 5 KBps respectively, then the Avg Network Utilization across all 3 VMs is 10 KBps. With the policy value set to 10% the dashboard will only display Idle VMs that have an Network Utilization of less than 1.00 KBps).
  2. Policy values are global and apply to all VMs in all environments.

Undersized VMs Policy

The Undersized VMs Policy includes four parameters. To be considered an Undersized VM, a VM must either meet both thresholds for CPU or both thresholds for vMem.

  • Average vCPU Utilization (default 75%) – This is the threshold for CPU Utilization, averaged over the time frame being analyzed (1 day, 7 days, or 30 days).
  • Maximum vCPU Utilization (default 95%) – This is the threshold for the maximum CPU Utilization experienced during the time frame being analyzed (1 day, 7 days, or 30 days).
  • Average vMemory Utilization (default 75%) – This is the threshold for vMem Utilization, averaged over the time frame being analyzed (1 day, 7 days, or 30 days).
  • Maximum vMemory Utilization (default 95%) – This is the threshold for the maximum vMem Utilization experienced during the time frame being analyzed (1 day, 7 days, or 30 days).

Oversized VMs Policy

The Oversized VMs Policy includes four parameters. To be considered an Oversized VM, a VM must either meet both minimum requirements for CPU or both for vMemory Usage.

  • Average CPU Usage (default 50%) – This is the minimum requirement for CPU Usage, averaged over the time frame being analyzed (1 day, 7 days, or 30 days).
  • Maximum Average CPU Usage (default 80%) – This is the minimum requirement for the maximum average CPU Usage experienced during the time frame being analyzed (1 day, 7 days, or 30 days).
  • Average vMemory Usage (default 50%) – This is the minimum requirement for vMemory Usage, averaged over the time frame being analyzed (1 day, 7 days, or 30 days).
  • Maximum Average vMemory Usage (default 80%) – This is the minimum requirement for the maximum average vMemory Usage experienced during the time frame being analyzed (1 day, 7 days, or 30 days).

vMemory Utilization vs Usage (for VMware)

To determine the Oversized VMs, SIOS iQ analyzes vMemory usage vs the utilization. vMemory usage is what VMware refers to as memory consumed.

Virtual machine: Amount of guest physical memory consumed by the virtual machine for guest memory. Consumed memory does not include overhead memory. It includes shared memory and memory that might be reserved, but not actually used. Use this metric for charge-back purposes.

vm consumed memory = memory granted – memory saved due to memory sharing

Here is how you can get high vMemory consumed when utilization is low. Utilization can be misleading at times because it only shows how much memory is being utilized of the amount granted. However, there is memory sharing that takes place, driven at the host level where memory can be used by another VM on the host. So, while the utilization can be low, the consumed could be up to 100%.

For example:

  • VM1 has 4GB granted
  • VM1’s utilization is 20% = 0.8GB

The host on which VM1 resides, due to potential over-provisioning, reallocated 3.2 GB (ballooning) to another VM (VM2 for example). Therefore, 4GB (granted) – 3.2 GB (that was ballooned to another VM) = 0.8 GB. While VM1’s utilization is 20%, total memory consumption is at 100%. Since VM1 consumed the additional memory that was available to it, the host managed that by ballooning it from other resources or swapping to disk.

Storage Acceleration

This section contains the parameters (Read Ratio, Cache Hit Ratio) that help define candidates for Storage Acceleration. A user may modify the settings to custom values or reset them to the default values. Storage Acceleration candidates are identified in the Storage Acceleration Dashboard, when all policy values are met or exceeded within the specified time period. Policy values are global and apply to all VMs in all environments.

  • Read Ratio (default = 70%)
  • Cache Hit Ratio (default = 50%)

Machine Learning

The Sensitivity setting provides the ability to control the sensitivity of the Machine Learning algorithms responsible for Performance Root Cause analysis. Of the three available settings – Low, Medium, and High – the High setting enables Performance Root Cause analysis based solely upon Learned Behavior. The default setting for Sensitivity is Low, which shifts the focus to more extreme behavioral changes, enabling detection of more severe issues.

  • Sensitivity (default = Low)

Capacity Forecasting

You may view or edit the Capacity Forecasting policy, which provides the ability to adjust the parameters of the capacity forecasting algorithm including the severity thresholds that determine related Performance Root Cause issues. An issue will be generated for a given severity level — Critical, Warning, or Informational — if the specified capacity usage percentage (Percent of Capacity Used) is forecast to be reached within the corresponding number of days designated by the policy.

  • Critical Number of Days (default = 30)
  • Warning Number of Days (default = 60)
  • Information Number of Days (default = 90)
  • Percent of Capacity Used (default = 100%)

Feedback

Was this helpful?

Yes No
You indicated this topic was not helpful to you ...
Could you please leave a comment telling us why? Thank you!
Thanks for your feedback.

Post your comment on this topic.

Post Comment