Skip to main content

Creating an event autotagger

Following analysis of the temperature data in the notebook, you determine whether additional robots have sensors impacted by temperature fluctuations.

Model-Prime Event Autotagging allows you to use log data to define events, to tag the events, and to search for the events.

Autotagging Overview

A tagger can be invoked in one of three ways:

  1. By testing the tagger,
  2. By invoking a tagger backfill against a set of logs that have already been ingested into the Model-Prime platform, and
  3. By enabling autotagging 'normfill' for all new logs that are ingested into the Model-Prime platform.

Creating an autotagger requires three steps:

  1. Defining variables: a variable represents the time series data that will be evaluated to determine whether an event occurred.
  2. Defining predicates: a predicate operates on the variable(s) you defined to determine whether they meet the required conditions of the event.
  3. Defining outputs: an output writes a new metadata value to a log to indicate that an event occurred.

Optionally, you can also define gating criteria to specify which logs should be processed by the tagger when normfill is enabled.

In this scenario, you will create a tagger that identifies events in which the temperature is greater than 28 degrees in the set of logs that have been provided.