Analysis - Editor

Overview

The Analysis Editor defines the metric and categorical dataset for an analysis along with sharing permissions and the datasources that prefer the analysis when drilling down.

Building an Analysis Dataset

Data selection is a familiar concept throughout Skylight Analytics. The standard selection of metric and categorical data is present here as well, with a few nuances distinct from Dashboards.

In particular, metrics and categories can be disabled initially. Deferring the selection of an expensive computation (e.g. Group by "Client IP") can significantly improve initial performance while still allowing the easy selection of the category when the dataset has been narrowed in scope.

Analysis also limits data selection according to analysis data types so as to form relevant associations between metrics and categories and provide required categories. When drilling down, an alernative analysis can be selected from analyses sharing the same data type, allowing the transfer of metrics and filters between analyses.

At the moment, there are six available types:

• Assets (also requires selection of a specific Asset type)
• Capture
• Session
• Alerts – Capture – Application: Application-based Capture Alerts
• Alerts – Capture – Zone: Zone-based Capture Alerts
• Alerts – Session: Session-based Alerts

Adding Data

Metric data can be selected from the Data Dictionary and customized using the standard controls.

You can add color rules to metrics to reflect, for example, good or bad ranges. Click the ellipsis (...), select Add color rules, and then set the parameters under Value and Shown as to define the condition and color. Repeat this step as needed.

Categorical data can also be selected from the Data Dictionary and has some limited customization, such as the selection of a Zone Level for Capture Zone categories. To allow proper function of the Analysis Widgets, required categories are listed but cannot be deleted or disabled.

The default Limit of 1000 rows can also be changed; a lower limit will result in faster performance but potentially limited context for correlations, while a higher limit has slower performance but may provide better overall context.

Edits can be rolled back at any point via X (Cancel).

Applying Changes

Save the analysis to persist changes permanently or Apply changes to re-compute the analysis dataset without persisting the changes.

Keep in mind that for an analysis shared with other users Save will impact everyone with access to the analysis. Save will also persist Filter and Widget states, such as the following:

• Filters currently applied to the Analysis
• Rearrange/maximize/hide widgets
• Rearrange/hide distribution and table columns

Sharing an Analysis

Analyses have sharing permssions identical to Dashboards. A new Analysis starts in a Private state accessible only to the author and may be shared with Everyone or specific User Groups as:

• Viewers: Read permissions (data selection is still available)
• Editors: Edit/save permissions
• Managers: Full permissions to Edit/Delete/Rename/Share

The Sharing dialog also allows Managers to review or audit current permissions.

Preferred Analyses for Datasources

When drilling down on a query/context, Skylight Analytics will attempt to select the most relevant Analysis according to the query/context Datasources.

For example, drilling down on "Delay (Max)" for a "TWAMP-SF" datasource will attempt to resolve a relevant analysis using the following strategy:

1. Find all Analyses with "TWAMP-SF" in the list of associated datasources, i.e. Default analysis template for: "TWAMP-SF".
2. Prefer Analyses created by the current user.
3. Prefer Analyses with more associated datasources matching the query/context datasource(s).

If no relevant Analysis is found, an empty Analysis will be created with the analysis data type associated with the query/context. This generated Analysis belongs to the current user and can be saved in order to provide a future default for the given datasource(s).

Datasource associations can be assigned via the Editor.