Skip to content

Latest commit

 

History

History
36 lines (23 loc) · 2.01 KB

README.md

File metadata and controls

36 lines (23 loc) · 2.01 KB
Raw

legacy vis analyzer

This package has two functionalities:

  • Analyzes the usage of legacy visualization types in beats and integrations dashboard and indexes them into an Elasticsearch instance.
  • Inlines the visualizations in a given directory as by values panels

Prep

  • Have locally installed node (tested with 16.14.2)
  • Init submodules using git submodule update --init --recursive
  • Install dependencies using yarn

Legacy vis analyzer usage

  • Run ES="<elasticsearch connection string>" node index.js
  • Import dataview.ndjson to have a bunch of runtime fields analyzing the structure

Code analyzer usage

  • Run ES="<elasticsearch connection string>" node index_kibana.js

Inliner usage

The inliner takes all "by reference" visualizations in a given directory, uses the provided running Kibana instance to migrate them to the latest version, migrates the dashboard saved object as well, then transforms the by-reference visualizations into by-value panels, deletes the visualization json files and updates the dashboard json files.

Important notes:

  • Using the inliner script will make the dashboards incompatible with earlier versions of the stack - e.g. if it has been ran with a stack version 8.2, then the new dashboard json files will only work on version 8.2 and newer

  • For old dashboards (prior to 7.10), some "agg based" visualizations might break if a target version of 7.17 or 8.0 is used. In these cases, please use at least a stack version of 8.1

  • Run KIBANA="<kibana connection string>" node inline.js <path to kibana folder> (e.g. ./integrations/packages/system/kibana/)

    • The kibana connection string has to include the password (for instances with security enabled) and the base path (for instances with configured base path), for example KIBANA="http://elastic:changeme@localhost:5901/mgp"

  • Review changes in submodule repo

    • This review should include loading the dashboard into an instance with data to make sure everything is displayed properly

  • If everything works fine, create PR