Understanding Capacity Views and Reports in VMware vRealize Operations Manager 5.x
search cancel

Understanding Capacity Views and Reports in VMware vRealize Operations Manager 5.x

book

Article ID: 343920

calendar_today

Updated On:

Products

VMware Aria Suite

Issue/Introduction

VMware vRealize Operations Manager has numerous factors that can affect its capacity planning components. This guide is a reference for how the components affect each other. This article provides the information needed to tune the dashboards, planning views and reports to provide better information about the environment.


Environment

VMware vCenter Operations Manager (vApp) 5.7.x
VMware vCenter Operations Manager (vApp) 5.8.x

Resolution

One of the important things to note about the dashboards is that the Risk and efficiency columns do not update in real time. They are calculated on a nightly basis. Any changes you make as we go on will not be reflected on the Dashboard tab until 1:00 AM. However, the dashboards and reports base their information on the Views under the Planning tab. Refer to them in order to see the changes you make in real time.


One of the instructive places to look for things that will affect the badge values and planning views is the policy configuration. Unless you have custom policy defined, the default policy can be edited. The default policy is safe to change, as it can be re-created if necessary.

Although the threshold values in section 2 are important, they are specific to every environment and fairly self-explanatory, so they will not be covered here. Section 3 has larger implications for what will appear in the planning views, however, so we will focus more on that.


Under 3a, it is common to change the following from the defaults:
  • Capacity Remaining based on Physical to Usable capacity
    • This means that capacity for hosts will be calculated based on their actual usage, rather than what the VMs have requested for hardware. This will be a more realistic scoring method for host density and overallocation so that you can get the most out of your hardware.
  • Remove all Allocation selections from the Compute Resource table
    • Similarly, this will determine what attributes of the VMs will be used in the above scoring. Allocation is very conservative, whereas demand is a very realistic representation of what is being consumed by the VMs.
  • Remove Network and Disk I/O from by Demand
    • Operations Manager does not have an API to discover the actual maximum hardware I/O rates, so it instead uses observed maximum, which is often substantially lower than your real maximums. Removing them from the Compute Resource table will reduce their impact on capacity calculations, which can sometimes be confusing if they are included.

Under Configuration > Display Settings, we come to some settings that will affect dashboards and reporting in a large way. Here, we define the rules for how things are averaged, over what range, and what the reports will show. The interval is the smallest metric size used for the planning views, dashboards, and reports.

Each View in the Views Gallery has a type. Some will be Trend
and some won't be. The settings are applicable only to those Views that are not Trend Views. For example, Summary, List and Distribution Views. When you pull up any Summary, List or Distribution Views, these Non-Trend Views settings would be applicable to the data the View is presenting.

Example Distribution View for cluster: Common Virtual Machine Configurations


Example Summary View for a cluster: Capacity Risk Details


Everything in these views is affected by Non-Trend Views settings with special handling for anything that is time remaining related metrics. For time remaining metrics in Non-Trend Views (if they are found), the interval to use is based on the Non-Trend Views but the number of data points would be based on Trend & Forecast setting. See the circled settings in this screenshot:



The rationale is that anything that is time remaining is based on the forecast of historical data trend. Capacity (not time) remaining is always based on an aggregation of historical data using Non-Trend Views settings.