I require help on optimising kibana dashboards. I got very slow loading on kibana dashboards. most of the time it's over 5 minutes.
Below graphs shows some information extracted from monitoring tools. When I am trying to load a dashboard with 10 visualizations it take around 4 seconds to process on elasticsearch server but its more than 100 seconds from kibana to load the dashboards.
3 node es cluster
1 node kibana on another node
If I recall correctly, the search latency shown for Elasticsearch is the latency per shard. As I assume you are querying more than one shard the full request latency can be a lot longer, which is shown in the latency data from Kibana, which shows t he full request including all aggregations against all shards.
How many indices and shards are you querying? How much data do you have in these indices?
What is the specification of the nodes in the cluster? What type of storage are they using?
Running dashboards with a lot of visualisations can cause a lot of disk I/O, especially if you are querying a large number of shards. Have you looked at what disk I/O looks like on the nodes while you are querying, e.g. using
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.