Slow datatable visualisation with multiple columns


#1

Hi,

Kibana version:6.4.2

I created a datatable vizualisation with 5 columns (4 + count column) but it takes 10+ seconds to load. Looking at the query details I saw it only takes around 100ms so the issue should not be on elasticsearch side.

There is only 8k documents in my elasticsearch db.

How is it possible to improve the performance ?


(Sjaak) #2

How many rows are in your table?

My understanding is that the search results will be sent to the browser as one big json and that can take a long time to process as its all single threaded. Take a look at the cpu usage on your client, you'll probably see your browser using up 25 ~ 30% while the visualization is loading.

I had the same issue and changed some of my tables from using @timestamp as a term to a data histogram to limit the amount of rows needed to be rendered.


#3

There is 50 rows displayed in my visualization and only around 8000 entries in elasticsearch.
The issue is exactly as you stated, the cpu usage goes up when rendering. However I it takes way too long to display this tiny amount of data (and the count value is always at 1, so there is no need for any aggregation here)
Unfortunately in my case I can't rely on the @timestamp key nor any other to filter out some data.

What I'm try to achieve is extremely close the the search panel, but you can't sort strings nor customize the headers in this type of panels.

Is there a way to achieve this with a correct rendering time ?


(Sjaak) #4

Most likely its the client side rendering that takes up most of the time so you could try on a different pc with better single thread performance and see if that speeds things up. You could also go into your visualization and delete columns one by one to check if there is one that somehow takes and inappropriately long time to render.


(system) #5

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.