Kibana5 load balancing for HTTP requests

Hi everybody,
I would like to know the best possible way to implement the load balancing for Kibana.
actually I have 3 nodes in a cluster; every node is master and data. I want to install kibana on another node (the fourth) using the elasticsearch coordinating only role.
can someone exaplain me what should I type in the field "transport.host" of elasticsearch.yml file?

using this role, which is the component that will balance the http requests? the local elasticsearch node, isn't it?
is there something else that I should consider?

thanks

a couple of clarifications here, to make sure we are on the same page

First, there is no elasticsearch coordinating only role, it is the whole node being coordinating only. Reason for this, that it can neither hold data nor become master.

Second, the transport.host setting has nothing to do with it. The documentation about coordinating only nodes shows the required settings.

If you have three nodes in a cluster that all have the same roles, you could point your kibana instance to just one of them. Or you get a fourth coordinating node up and running and point kibana to that one. This coordinating only node would then do the balancing by redirecting all the incoming requests from kibana inside the cluster.

Hope this helps!

--Alex

1 Like

First of all thanks to your answer.
so, Kibana will aim to localhost:9200 and then, local Elasticsearch will know every other cluster's nodes and will redirect to one of them (in the theory, to the best one) my HTTP Kibana request?

yes, that is what a coordinating only node basically does. That node knows the cluster topology and can forward all requests directly.

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