upmc-enterprises
How to expose resources created with the `operator` helm chart
How can I expose resources created with the operator helm charts (on AWS) ?
kubectl expose deployment.extensions/kibana-database-cluster \
--type=LoadBalancer --name=kibana-access
kubectl expose service/elasticsearch-database-cluster \
--type=LoadBalancer --name=elasticsearch-access
Services and load balancers has been created but do not seem to work properly. Connection just hangs and brakes on timeout then.
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
elasticsearch-access LoadBalancer 100.65.247.239 af95c1a0046df... 9200:30255/TCP 33m
kibana-access LoadBalancer 100.71.244.196 af913a79646de... 5601:31385/TCP 40m
What could be wrong here? Am I missing something? Thank you.
First of all i'd recommend exposing these type of services through Ingress. Which takes some (undocumented) effort, but it's probably different for each implementation anyway.
But if you do want to directly expose your ES/kibana access (bad idea) check if your expose rules actually selected the right backends. kubectl describe and kubectl get ep should get you somewhere.
the Elasticsearch chart doesn't support ingress creation, a PR adding this feature are very welcome
As @pieterlange mentioned exposing services without any authentication method is a really bad idea.
@komljen exposing an ingress using internal load balancer + oauth proxy for services like kibana is a must. A note in the helm docs explaining how to do this will be required
I wrote a blog post on this topic a few days ago https://akomljen.com/protect-kubernetes-external-endpoints-with-oauth2-proxy/. We could add the same instructions in some docs in this repo for Kibana deployed with the operator.