Skip to content

Latest commit

 

History

History
67 lines (57 loc) · 2.85 KB

endpoint-settings.md

File metadata and controls

67 lines (57 loc) · 2.85 KB
Raw

Endpoint setting for Elasticsearch

When you deploy Elasticsearch cluster, you need to have a public endpoint for user and remote application access (not hosted on the same kubernetes cluster). To do that, you can use:

  • Ingress endpoint: Access to Elasticsearch from standard kubernetes ingress (reverse proxy)
  • Load balancer endpoint: Access to Elasticsearch from kubernetes Load balancer (not supported on all kubernetes cluster)

Ingress endpoint

You can use the following settings:

  • enabled (boolean): Set to true to enable ingress endpoint. Default to false
  • targetNodeGroupName (string): You can retrive endpoint traffic to particular node group, like client node group. Default it will load balance through all nodes
  • host (string / required): The hostname to access on Elasticsearch API. No default value
  • secretRef (object): If you set it, it will enable https. If secret exist, it will use the certificate provided, else it will used the default ingress certificate
    • name (string / required): the secret name
  • labels (map of string): The list of labels to add on ingress. Default to empty
  • annotations (map of string): The list of annotations to add on ingress. Default to empty
  • ingressSpec (object): You can set any other ingress properties. Read the official doc to know the properties

elasticsearch.yaml:

apiVersion: elasticsearch.k8s.webcenter.fr/v1
kind: Elasticsearch
metadata:
  name: elasticsearch
  namespace: cluster-dev
spec:
  endpoint:
    ingress:
      annotations:
        nginx.ingress.kubernetes.io/proxy-body-size: 512M
        nginx.ingress.kubernetes.io/proxy-connect-timeout: '600'
        nginx.ingress.kubernetes.io/proxy-read-timeout: '600'
        nginx.ingress.kubernetes.io/proxy-send-timeout: '600'
        nginx.ingress.kubernetes.io/ssl-redirect: 'true'
      enabled: true
      host: elasticsearch-cluster-prd.domain.local
      secretRef:
        name: es-tls
      ingressSpec:
        ingressClassName: nginx-example
      targetNodeGroupName: client

Load balancer endpoint

The kubernertes cluster must be support load balancer service. On cloud provider it can work out of the box. Is not the case on onprmise cluster. You need to deploy extra stack like metalLB.

You can use the following settings:

  • enabled (bool): Set to true to enable load balancer. Default to false.
  • targetNodeGroupName (string): You can retrive endpoint traffic to particular node group, like client node group.

To get Load balancer IP, you need to get the IP on status service.

elasticsearch.yaml:

apiVersion: elasticsearch.k8s.webcenter.fr/v1
kind: Elasticsearch
metadata:
  name: elasticsearch
  namespace: cluster-dev
spec:
  endpoint:
    loadBalancer:
      enabled: true
      targetNodeGroupName: client