What is the most efficient way to implement visualizations in dashboard

I can have 1 visualization split on category showing 3 graphs for CPU for example (3 different categories of server)

Or I can have 3 separate visualizations (1 per category) which looks neater as the legend is on each graph and less cluttered.

If they all use the same underlying search, do I have to pull the data back 3 times from elasticsearch for the 3 separate visualizations, or since they are all in the same dashboard, does the search get executed only once ?

Many Thanks

Mick

Great question. The Kibana front-end registers all the queries that it needs to render a page, and sends them to Elasticsearch as an _msearch.

The NodeJS server for Kibana proxies requests from the browser to the Elasticsearch connection, so you can actually see the msearch query being sent if you watch the network traffic from the browser.

So whether you have a single visualization or multiple, getting the data is a single query to Elasticsearch.

That explains it very well, and thats what I was hoping - excellent :slight_smile:

Many Thanks Tim,

Mick