Insights

Discover signals from automated analysis of log data

An Insight is a human readable explanation and quantification of behavioral signals generated from unsupervised analysis of enriched logs as part of the Data Pipeline. Insights represent signals that the user can triage and refine through the Explorer or inspection of Logs and track as Metrics or Segments.

Insights are one of three types:

  • Temporal Insights: Detected changes in the behavior of one or more Columns of log data.

  • Segment Insights: Detected clusters related to filters on Columns of log data that correspond to unique behavior.

  • Outlier Insights: Specific instances of log data that deviate from expected behavior.

Temporal Insights

Temporal Insights correspond to changes or shifts in behavior related to one or more Columns (including Metrics). It is defined by a time split within a time window and an impacted Column. The time split represents a before and after in which the temporal insight was observed.

To investigate a temporal insight:

To track a temporal insight:

  • Create a Metric: Capture the underlying behavior that led to the change in a custom Metric.

Segment Insights

Segment Insights correspond to patterns in behavior related to one or more Columns (including Metrics). A segment is defined by a filter on one or more Columns that bifurcate the log data.

To investigate a segment insight:

  • View in Explorer: Takes you to Segment Comparison in Explorer to view differences in Columns (including Metrics) between the segment and the rest of the log data.

  • View Logs: Takes you to the Logs page with the Segment filter applied to view explicit examples of logs within the Segmnet.

To track a segment insight:

  • Create a Segment: Capture the filter representing this Segment for future analysis and display on the Segments Dashboard.

  • Create a Metric: Capture the underlying behavior that represents this Segment in a custom Metric.

Outlier Insights

Outlier Insights correspond to specific examples of large deviations in behavior related to one or more Metrics. It is defined by a specific set of logs that represent this behavior.

To investigate a outlier insight:

  • View Logs: Takes you to the Logs page with a filter applied to view explicit examples of outlier logs.

To Track a temporal insight:

  • Create a Metric: Capture the underlying behavior that led to the change in a custom Metric.

Was this helpful?