I'm trying to deploy Elastic and Kibana in a Kubernetes cluster.
I have installed Elastic using Helm chart :
helm repo add elastic https://helm.elastic.co
helm repo update
helm install stable/elasticsearch --namespace elastic --name elasticsearch --set imageTag=6.5.4
And Kibana using Helm chart :
helm install elastic/kibana --namespace elastic --name kibana --set imageTag=6.5.4,elasticsearchURL=http://elasticsearch-client.elastic.svc.cluster.local:9200
I've checked from my Kibana pod, and this URL is reachable and produce the following result
curl -v http://elasticsearch-client:9200
* About to connect() to elasticsearch-client port 9200 (#0)
* Trying 10.19.251.82...
* Connected to elasticsearch-client (10.19.251.82) port 9200 (#0)
> GET / HTTP/1.1
> User-Agent: curl/7.29.0
> Host: elasticsearch-client:9200
> Accept: */*
>
< HTTP/1.1 200 OK
< content-type: application/json; charset=UTF-8
< content-length: 519
<
{
"name" : "elasticsearch-client-8666954ffb-kthcx",
"cluster_name" : "elasticsearch",
"cluster_uuid" : "-MT_zbKySiad0jDJVc1ViQ",
"version" : {
"number" : "6.5.4",
"build_flavor" : "oss",
"build_type" : "tar",
"build_hash" : "d2ef93d",
"build_date" : "2018-12-17T21:17:40.758843Z",
"build_snapshot" : false,
"lucene_version" : "7.5.0",
"minimum_wire_compatibility_version" : "5.6.0",
"minimum_index_compatibility_version" : "5.0.0"
},
"tagline" : "You Know, for Search"
}
The command line used in the Kibana pod to start (generated by the helm chart) is
/usr/share/kibana/bin/../node/bin/node --no-warnings /usr/share/kibana/bin/../src/cli --cpu.cgroup.path.override=/ --cpuacct.cgroup.path.override=/ --elasticsearch.url=http://elasticsearch-client:9200
So it seems the Elastic cluster url is the right one, and reachable.
However, when I show the UI in my browser, I get the following page
So to sum up, both versions are identical :
ElasticSearch url is correct, but Kibana don't want to access ElasticSearch
I tried this myself and there's something with the Kibana docker image and/or Helm chart on how the parameter is passed into Kibana. Basically, the command line shows:
--elasticsearch.url=http://elasticsearch-client.elastic.svc.cluster.local:9200
But if you shell into the container/pod you see that Kibana command line expect something different for the elasticsearch URL (-e, --elasticsearch <uri>
):
$ /usr/share/kibana/bin/kibana --help
Usage: bin/kibana [command=serve] [options]
Kibana is an open source (Apache Licensed), browser based analytics and search dashboard for Elasticsearch.
Commands:
serve [options] Run the kibana server
help <command> Get the help for a specific command
"serve" Options:
-h, --help output usage information
-e, --elasticsearch <uri> Elasticsearch instance
-c, --config <path> Path to the config file, can be changed with the CONFIG_PATH environment variable as well. Use multiple --config args to include multiple config files.
-p, --port <port> The port to bind to
-q, --quiet Prevent all logging except errors
-Q, --silent Prevent all logging
--verbose Turns on verbose logging
-H, --host <host> The host to bind to
-l, --log-file <path> The file to log to
--plugin-dir <path> A path to scan for plugins, this can be specified multiple times to specify multiple directories
--plugin-path <path> A path to a plugin which should be included by the server, this can be specified multiple times to specify multiple paths
--plugins <path> an alias for --plugin-dir
--optimize Optimize and then stop the server
So, something is not translating the elasticsearch URL correctly.
It seems like the default is localhost:9200
so you could try a sidecar container in your kibana deployment so that forwards everything on port localhost:9200
to elasticsearch-client.elastic.svc.cluster.local:9200
. Perhaps following this
I think you're using the OSS Elasticsearch distribution and the Non-OSS Kibana package.
Can you try with docker.elastic.co/kibana/kibana-oss:6.5.4 ?