Viewing correlated metrics
For example, if there is an increase in the time required for loan approval (Loan Approval Time) and a proportional (direct or inverse) change in memory usage (Memory Used), they can be considered as correlated metrics.
To view correlated metrics
- Select any metric that is collected from any source to view its details.
- Click the action menu
next to the metric name and select Correlate.
A correlation analysis finds and displays metrics that correlate with the selected metric.
If a metric correlation analysis takes more than several seconds, a dialog box opens that enables you to run the correlation analysis in the background. - To view details about previous correlation analyses, click the action menu
next to the metric name and select Recent Correlation Analyses.
Details are displayed for the five most recent correlation analyses that were run in the last seven days:- Time Started—Time and date that the correlation analysis started. Click the the start time to see the specific analysis.
- Status—Status of the analysis (In Progress, Finished)
- From and To—Time and date range of the analysis
Metric Correlation page
The correlation algorithm, by default, correlates the selected metric with all other metrics within the same application. If the metric is not associated to an application, it is correlated with all metrics in the account. The currently selected time range is used to perform the correlation.
The metric correlation results view shows a graph with the primary metric and up to five of the most highly correlated metrics to the primary metric. Additional correlated metrics are displayed in the table below, and are sorted to first display the metrics with the highest correlation. Select the check box corresponding to a metric to remove or add it to the graph.
After reviewing the initial correlation results, you can adjust the time range or the available filters on the left to run a new correlation analysis.
Parts of the Metric Correlation page
Click a section in the image to view details.
Charts and correlated metrics list (Section A)
The correlation algorithm, by default, correlates the selected metric with all other metrics within the same application. If the metric is not associated to an application, it is correlated with all metrics in the account. The currently selected time range is used to perform the correlation.
Below the chart are details of the correlated metrics. The first row displays the primary metric selected for correlation analysis. The metrics correlated to the primary metric are displayed in the following rows, with the highest correlation displayed first.
The following details are displayed below the chart:
- Metric: Name and units of the correlated metric
- Correlation: Correlation coefficient for the metrics.
The correlation calculated in TrueSight Intelligence is the Pearson Correlation Coefficient. It is reported as a value between -1 and +1. - Lead/Lag: Indicates if the correlation determined takes a time lead/lag, with respect to the primary metric, into consideration. For instance, if a metric leads by n number of minutes, it means that the leading metric is a probable cause. However, if a metric correlates with the primary metric with a lag, the primary metric is a probable cause.
- Source: Name of the source to which the metric belongs.
- Chart: Displays a thumbnail for the metric for the selected time period and provides an indicator of the color used for plotting the metric in the correlation chart. Click this thumbnail to display a larger chart for the metric.
Time range (Section B)
When you analyze the primary metric data series with each of the other metric data series, the time-series data is compared only within the selected time period. Change the time period of the analysis to see a potentially vast change in the results.