Performance management and capacity planning vary across organizations and environments. Because the demand for capacity fluctuates in each environment, the top contenders for priority often include high efficiency versus low risk of poor performance. To plan and manage your capacity needs and intelligently calculate the capacity of your resources, vRealize Operations Manager uses sophisticated models.
With the available capacity calculations, you can use various sophisticated models to produce practical correlations between objective measured metrics and subjective goals of acceptable performance and efficiency.
For example , the stress concept involves how high and how long the demand persists relative to the capacity available, and vRealize Operations Manager uses this value to measure the potential for performance problems. The higher the stress score, the worse the potential is for degraded performance on your objects. Depending on the configuration of the policy analysis settings for stress, a score of green might indicate 0–24 percent of stress. A score of red might indicate more than 50 percent of stress. With the five-minute data collections and the intelligent stress calculations, the system easily identifies periods of poor performance.
Demand drives stress. The system bases the calculations for right-sizing capacity on past demand. The goal of right-sizing is to produce a green stress level, marked by a green Stress badge.
Usable capacity is equal to the total capacity available minus any buffers that administrators or users defined. To measure the right-sized amounts of usable capacity, the capacity calculations use what is called a stress-free value. Using the demand, stress, and the stress-free value, vRealize Operations Manager calculates the right size.
The capacity analytics determine the actual and effective demand for resources based on having no contention. The calculations consider the capacity to be unlimited and free of contention for resources, which results in no stress on the available capacity. The result is called the stress-free demand or the stress-free value.
Where to Find Stress-Free Demand and Stress-Free Value
In some areas of the user interface, vRealize Operations Manager identifies capacity as Stress Free Demand, and in other areas it is identified as Stress Free Value. Both terms mean that the calculated capacity for an object is free from unacceptable levels of contention and stress, as defined in the policy for the Stress score.
Stress Free Demand appears in All Metrics, Views, and Reports.
In All Metrics, you can use the metric named Stress Free Demand to examine the CPU demand, disk space allocation and demand, memory consumed, and the vSphere configuration limit on an object. When you apply this metric to these resources, you can build a metric graph to display the stress-free demand for an object. The graph displays the high and low stress-free capacity values over time.
In Reports, you can use a view that includes the metric named Stress Free Demand to generate a report. The table in the report displays Stress Free Demand as the label. For example, this metric appears in the report named Cluster CPU Demand (%) Trend View.
Stress Free Value appears on thetab, and on the tab.
On the vSphere configuration limit. In this view, the table column name is Stress Free Value.tab, you can view the time remaining for CPU demand, memory consumed, disk space demand and allocation, and the
On the vSphere configuration limit.tab, the table column name is Stress Free Value. The tables display Stress Free Value as the calculated values for CPU demand, memory consumed, and the
Setting the Thresholds for the Stress Score
The analysis settings in the policy that you apply to your objects defines the thresholds for the stress score. The policy includes default settings for the stress score to be green, yellow, orange, or red. If the settings are too strict or loose for your environment, you can modify them.
To modify the stress score thresholds, edit the policy that applies to your objects, and click Analysis Settings. Select an object type and click the filter icon to display the policy analysis settings. In the Stress area, click the lock icon, expand Stress, and modify the stress thresholds.
In the analysis stress settings, vRealize Operations Manager uses the selected resources, such as Memory Demand, CPU Demand, and vSphere Configuration Limit to calculate the stress score.
You can set the stress thresholds to your own values, or turn them off. To change a stress score threshold, click and drag an icon along the slider. To remove a scoring range, such as the default range of 35–49 identified by orange, double-click an icon to disable the range.
Demand Exceeds is a percentage of capacity. Capacity is also called provisioned capacity. To change the stress threshold for a resource, double-click the Demand Exceeds percentage, and enter the desired value. This value defines the point at which vRealize Operations Manager considers the percentage of demand to be stress. For example, to change the stress threshold for Memory Demand, double-click the current percentage, such as 70.0 % of capacity, and enter the new percentage of demand to exceed for vRealize Operations Manager to identify stress.
For each resource, you can change the sliding analysis window value to include the entire range, and set the peak value to a different time depending on how you need vRealize Operations Manager to derive the stress score.
More About the Stress Score
vRealize Operations Manager calculates the stress zone and stress score for you. The following explanations cover typical scenarios where Demand does not exceed Capacity.
To determine the stress on an object for a specific time period, you can examine the demand curve to determine how much of the stress zone the demand occupies. The stress zone is typically where demand exceeds 70 percent of the total capacity. For example, stress occurs when CPU demand, memory demand, or memory consumed exceeds 70 percent of the capacity.
In a 60-minute peak period, vRealize Operations Manager bases the Stress score calculation on the following variables:
Stress threshold, which is the Demand Exceeds setting
Stress score threshold, which determines the color of the Stress badge
Time range, as in 30 days of analysis
Peak detection window, which is the 60-minute peak setting that you can adjust to either a non-zero number of minutes or the entire range.
When the demand exceeds 70 percent, that data point in time is in the Stress zone.
In the policy stress analysis settings, to examine an example graph used to calculate stress, click What is stress?.
Another example to explain the calculation used for CPU stress is shown here.
With a peak detection window size of 60 minutes, vRealize Operations Manager calculates the CPU stress score. It uses the area under the demand curve and above the stress threshold line as a percentage of the area covered by the total capacity curve.
Using time stamps of t1 and t2 to identify a 60-minute window in the last 30 days, the stress score depends on demand, stress threshold, and total capacity over time.
Maximum((Demand - Stress Threshold) ÷ (Total Capacity - Stress Threshold))
This equation applies to the stress calculations for each resource, such as memory demand, memory consumed, and CPU demand.
If Total Capacity varies during the time range being considered, Stress Threshold must also become variable, because (Stress Threshold) = (Stress Threshold in %) × (Total Capacity).
Since (Total Capacity) can be a different value at a different time, as identified by t, then “Stress Threshold”(t) = “Stress Threshold in %” × “Total Capacity”(t).
As a result, the Stress score is the highest aggregate of demand that exceeds 70 percent of capacity, as a percentage of the aggregate of capacity within any contiguous interval of 60 minutes in the last 30 days. The formula for the score is as follows:
Maximum((Demand(t1, t2) - “Stress Threshold”(t1, t2)) ÷ (“Total Capacity”(t1, t2) – “Stress Threshold”(t1, t2)))
t1 and t2 are time stamps in the time continuum within the last 30 days.
t1 < t2
t2 - t1 = 60 minutes
Demand(t1, t2) is the demand curve between time t1 and t2.
“Stress Threshold”(t1, t2) is the stress threshold curve (as absolute values) between time t1 and t2.
“Total Capacity”(t1, t2) is the capacity threshold curve between time t1 and t2.
vRealize Operations Manager calculates the aggregate during a contiguous time interval of 60 minutes in the last 30 days. The Stress score is the percentage of aggregate capacity in the same contiguous time interval of 60 minutes. An acceptable score yields a green Stress badge.
To view the Stress zone for an object, click. Then, examine the Stress breakdown areas for CPU and memory, the Stress Zone column in the table, and the graph of actual demand.
By calculating the stress score, vRealize Operations Manager provides an intelligent way to evaluate peaks and fluctuations of the capacity of your objects over time.