This is a new custom visualization (for Kibana version 4.5.1) that allows displaying a dendrogram chart to show and index or classes, or some other hierarchical structure.
If you really liked this and feel like making a donation :
If you need a modified version, or one that works on a previous (or future) Kibana release, please let me know.
Example of how it looks:
You can just copy the dendrogram_vis folder into KIBANA_HOME/src/plugins) and run kibana (with --dev parameter).
You will have to re-bundle, if you just start the service or run in non-dev mode.
In order to work this plugins needs a simple Schema configuration:
- One Metric Parameter (whatever, really optional)
- One aggregation (bucket) to take the structure from.
(See flare.csv as an example)
In this project there is a folder (dendrogram_data) with a script in python to get a CSV into the elasticsearch server.
(The python script requires elasticsearch and csv packages).
The index on the elasticsearch should have a field called "id", that contains (for instance):
parent
parent.child1
parent.child2
parent.child1.child1
...
and so on.
To make it work you have to have the latest D3 distribution (Which up to today is #4).
The problem is that Kibana 5.5 comes with version 3, so you have to:
(whils on the dendrogram_vis folder)
npm install d3 d3-tree
You MAY encounter an issue with xmlhttprequest (not finding a module called child_process), which I have fixed by adding:
"browser" : {
"child_process" : false
}
To the xmlhttprequest package.json file (that got installed by npm on the node_modules folder).
As long as you don't actually USE xmlhttprequest, from the visualization, it should be ok.
- Add filtering from the nodes (clicking a node adds/remove filters)
- Zooming and Panning
- Colors (or icons) change based on data