Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
 

Data Summarization Overview

Paragon Insights provide a way to store raw data using summarization profiles to reduce the disk space and improve performance of time series database (TSDB).

Paragon Insights collects data from devices by using push or pull data collection ingest methods. You can create rules or use the available pre-defined rules to determine how and when data is collected. The telemetry data can be summarized as a function of time or when a change occurs.

For time-based data summarization, the raw data points are grouped together into user-defined time spans, and each group of data points is summarized into one data point using aggregate functions.

Paragon Insights also supports data rollup summarization. Data rollup summarization helps you to summarize field-level data. Field-level data is processed data that provides information on network devices and its components, and is stored in fields in the TSDB. A field is a single piece of information that forms a record in a database. In TSDB, multiple fields of processed data make a record. Data rollup summarization enables efficient data storage and also ensures retaining of data for a longer duration.

Table 1 provides a list of the supported data summarization algorithms and a description of their output:

Table 1: Descriptions of the Data Summarization Algorithms

Algorithm

Description of output

Latest

Value of the last data point collected within the time span.

Count

Total number of data points collected within the time span.

Mean

Average value of the data points collected within the time span.

Min

Minimum value of the data points collected within the time span.

Max

Maximum value of the data points within the time span.

On-change

Value of the data point whenever the value is different from the previous data point (occurs independently from the user-defined time span).

Stddev

Standard deviation of the data points collected within the time span.

Sum

Sum of the data points collected within the time span.

If no summarization algorithm is associated with the data, the following algorithms are used by default:

Data type

Data summarization algorithm

Float, integer, unsigned

Mean

Boolean, string

On-change

You can use data summarization profiles to apply specific summarization algorithms to raw data and field-level data collected by Paragon Insights for a specific device group:

These topics provide instructions on how to create a data summarization profile.