Browse Prior Art Database

Highlighting Outliers For Secondary Metrics

IP.com Disclosure Number: IPCOM000241694D
Publication Date: 2015-May-22
Document File: 4 page(s) / 88K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for highlighting outliers for secondary metrics using a pie chart. When an outlier metric is distinctly shown, the other metrics are dimmed in order to highlight the outlier.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 42% of the total text.

Page 01 of 4

Highlighting Outliers For Secondary Metrics

A compact graph showing multiple monitored metrics (e.g., a pie chart), often contains secondary metrics that are mainly important only when they are outliers. In Figure 1,

which is an example, for enterprise database activity time (i.e. how much time is spent in a database across various activities). Activities of primary importance in the illustrated instance are Structured Query Language (SQL) execution, input/output (I/O), and Lock waits. That is, SQL execution, input/output and lock waits are three primary activities. Slices are shown in the pie chart for each respective one of these primary activities. The size of each slice in the chart indicates the amount of time that an enterprise database spends doing each of these activities.

Activities of secondary importance are shown in the "other" category in Figure 1. In one example, there may be several secondary activities, one of which may be sort processing. This single slice represents the aggregated time spent on all the secondary activities, because these activities are usually relatively less important than the primary activities. In the illustrated instance, the combined time spent on the secondary activities is small, so the slice labeled "other" is small. This is not unusual, although it is not a requirement for the combined time for secondary activities to be small.

Figure 1: Activities in an example enterprise database system

The method disclosed herein shows a metric graph such as Figure 2 in a compact way,

wherein in a normal mode the secondary metrics are aggregated in a single slice (left hand pie chart of Figure 2) when no outliers are present among them, but when one of the secondary metrics becomes an outlier, the outlier metric is distinctly shown in a new slice (slice E in right hand pie chart of Figure 2), which is in addition to the slices A, B and C of the primary metrics and in addition to the slice of the other secondary metrics. When the outlier metric is distinctly shown, the other metrics are dimmed in order to highlight the outlier. Thus, in a default situation, when no outliers are present, the visual clutter is reduced; however, when an outlying condition occurs, it is easy to identify the outlier.

1


Page 02 of 4

An outlier may be defined by deviation from a normal amount of time spent doing an activity. When the enterprise database spends more time doing a secondary activity E that was previously included in the "other" category, the time spent doing E, e.g., sort processing, may exceed some threshold such that the pie chart changes to show a slice dedicated to activity E. Some implementations might set a simple threshold,

wherein activity time exceeds a predetermined, absolute amount of time. In other implementations, a dynamic threshold is provided. For a dynamic threshold, sampling determines averages and standard deviations of times spent doing secondary activities in predetermined intervals, so tha...