Kibana data table visualization is too slow in response

(Rajeshkumar) #1

I have 5 node Elasticsearch cluster 3 master node 1 data node and 1 Tribe node. I have installed ELK 5.2 version. My shards count is 2 and replica is 1. I have 32GB RAM and 1TB of hard disk space. My Index size is around 14GB. I have configured my kibana.yml to point both the tribe and master node

elasticsearch.tribe.url: "http://xxxx:9200"

elasticsearch.url: "http://xxxxx:9200"

Whenever I created a data-table visualization with date datatype as one of the column my search becoming too slow sometimes I am getting timeout error. When I remove the date type column form the visualization the search is better.

Is there any setting which I need to add in the config file or need to modify any elastic search mapping. Please advice me the configuration to overcome the issue.

Kibana visualization is too slow
(Nathan Reese) #2

Why are you doing a terms aggregation on a Date field? Your terms aggregation is asking for a very large number of terms. Using a Date Histogram aggregation will fix your performance problems.

(Rajeshkumar) #3

@Nathan_Reese When I use Date Histogram Yes its too faster but I couldn't see the minutes, sec and milliseconds information.

(Nathan Reese) #4

Why are you sub-aggregating the date histogram bucket with a terms aggregation on the same date field? Can you describe your use case and data set so I can get a better understanding of what you are trying to accomplish.

(Rajeshkumar) #5

Just to show you the difference I created this same field.

(system) #6

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