When I checked the templates of Kibana there is no statefulset.yaml file
I was just wondering how Kibana container deployment happening when there is no api version mentioned the way it is present in
/elasticsearch/templates/statefulset.yaml.so how the helm-charts were being used and decided which api version to be used for deployment
That means Elasticsearch always comes up with elasticsearch-master-0,elasticsearch-master-1,elasticsearch-master-2 as the Statefulset maintains sticky identity for each of their pods and elasticsearch-master-0 always remain as MASTER pod. Can't we do the same to Kibana if we replicate or scale the instances of Kibana with Master & Slave concept. I am just wondering why and why not for Kibana if the same we are doing for elasticsearch? I am just thinking what so special about Elasticsearch that same sort of features can't be extended to Kibana,
If any limitation for Kibana.
I am just thinking what so special about Elasticsearch that same sort of features can't be extended to Kibana
Elasticsearch is stateful application which manages data and need to handle clustering, this is why we need to use
StatefulSet K8S resource.
Kibana is a totally stateless application which doesn't manage any data, and doesn't have any Master & Slave concept.
If you want to scale Kibana, you just need to deploy new Kibana instances targetting the same Elasticsearch cluster and put them behind a Load Balancer.
This is why we can use
Deployment K8S resource for it which manage scaling by starting new pods and adding them to the K8S
If you need more details, I invite you to read Elasticsearch and Kibana documentation, as well as K8S documentation for
Deployment to understand the differences.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.