Questions about loadbalancing kibana to multiple ES nodes

Hi,

I am currently setting up my first multi node cluster of the stack.

How do I configure kibana correctly to loadbalance over the ES nodes?

In Kibana config I have the key elasticsearch.hosts.
Do I understand correctly kibana will distribute it's requests over all nodes if I place all cluster nodes here?
What if an additional node joins, does it need to be configured here too, or will kibana only need one available ES node and fetches cluster information (which nodes are available) from there?
Is it mandatory to list all nodes to get complete data or would 1 available ES node be enough?

I found this doc here:
https://www.elastic.co/guide/en/kibana/current/production.html#load-balancing

Where exactly is the benefit of running coordinating-only-nodes? Is it only to free up the coordinating resources from the other nodes or is it because the coordinating only node will discover all ES nodes automatically to loadbalance while kibana needs to be configured manually?

Thanks, Andreas

1 Like

Hey @asp,

The Coordinating-Only Node is effectively a load-balancer to multiple ES nodes. It is also responsible, as you noted, for offloading the creation of the final result set (see https://www.elastic.co/guide/en/elasticsearch/reference/7.1/modules-node.html#coordinating-node).

The benefit of having multiple ES nodes specified in your kibana.yml is for high availability. If you only have the single Coordinating Node specified, then Kibana will only function if that specific node is available. If that node falls out of the cluster, then Kibana will not be usable. Specifying multiple nodes here protects you from that scenario.

1 Like

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