We have a two data nodes, three masters, and one kibana. There is an azure load balancer on top of the data nodes. We are looking into load balancing across these nodes via Smart Load Balancer suggested in the documentation.
If we are understanding correctly we need to have the load balancer on the same vm as kibana and point kibana as well as all of our queries to this load balancer.
If that is the case here is our concern, and we're looking for suggestions or how to handle such an event, is that what happens in the event the vm holding kibana and the load balancer goes down(os updates, reboots, azure maint, etc)?
What is the best way to handle such an event? Or are we misunderstanding how to configure this properly?
Hey @AArban, you don't have to put the load balancer on the same VM as Kibana, and if you're going to deploy Kibana in a high-available manner, I'd advice against it. Kibana stores all of it's persistent data in Elasticsearch, so if Kibana is restarted/reprovisioned by Azure it won't lose any of your Dashboards/Visualizations, etc.
We do recommend install a coordinating Elasticsearch node on the same VM as Kibana, and having Kibana connect to this instance of Elasticsearch. Is this perhaps what you were thinking of in regard to having the load balancer on the same VM as Kibana?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.