Time Series uses aggregation techniques to display metrics across large time intervals when long period values are chosen for the charting action. The charting tool uses this aggregated data to represent one or more actual data values that are sampled across the time interval that is dictated by the selected period.
When you select a long time period value to chart, the interval is adjusted to display information in a more readable way. The bigger the interval, the more potential data that is sampled and used for the aggregated value. This value displays as a single plot point at the end of the time interval. When multiple sampled data points fall into an interval window, those points are aggregated using a calculation method of averaging or summing.
All storage metrics are defined to use an averaging method to calculate the aggregated data points. When multiple values are rolled up for a time interval or across multiple entities, the average value is represented on the Time Series charts. So, for long period time ranges, the displayed plot values could be aggregating one or more samples plotted at the end of the time interval.
Example 1 Storage Interval Aggregation
The selected chart period dictates that hourly intervals are plotted. For the time range 0100-0200, Volume ABC001 has two data samples for the selected metric (100, 150). For the metric chosen, these two values are averaged and represented as 125 on the chart.
Example 2 Storage Contributor Base Entity Aggregation
The user has accessed the TSF Contributors. The user unchecked the Volume Name in the Base Entity criteria to chart a metric across all volumes for a system.
For the time range 0100-0200, each volume has only one value. For ABC001=100, ABC002=200, ABC003=150, ABC004=50, these values are averaged to 125. If the user plots contributors, an individual contributor, such as ABC003, could plot above the chart for the combined result.
|
Copyright © 2013 CA Technologies.
All rights reserved.
|
|