We have a two machine setup, in which we will be installing whole ELK stack. The elastic search nodes will be working as a cluster.
My problem is, since Kibana can link only to a single Elastic node and not cluster, so if say es1 process goes down, and if the user is trying to access Kibana1, it will not be able to fetch the data, and he has to manually switch to Kibana2.
Is there any way to auto-switch the Kibana1 to say ES2 node? Or is it possible to solve this using some other technique. My only constraint is that we have only two machines, and we cannot increase it.
Use a client node on the same server as KB, that'll work.
But there is already a ES master/data node running on that server.
So basically the setup is as follows:
Machine1: ES DataNode, LS, KB
Machine2: ES DataNode, LS, KB
Is there any way Kibana "sees" all the nodes in ES cluster.. So that if one ES node goes down, it can fallback to the second ES node in round robin manner.
This is very basic requirement, why KIbana is not based on this strategy?