Below are instructions for setting up a simple Elastic + Kibana demo on DCOS
- DC/OS 1.12 EE
- X Masters
- Y Agents
Connect to your cluster using dcos cluster setup
like so:
dcos cluster setup https://<MASTER_NODE_IP_OR_ELB_ADDRESS>
If you don't already have the enterprise cli installed, do so.
dcos package install dcos-enterprise-cli --yes
dcos package install elastic --yes
You can monitor the Elastic deployment with the command below:
dcos elastic plan status deploy
First, SSH into your master node. This is where we will pull down our data and load it into Elastic.
dcos node ssh --master-proxy --leader
Next, we will pull down the sample data we will use.
NB: This is a simple dataset of US spending on hospital care by state from 2008-2014. This dataset was selected because health spending a universally understandable metric and lends itself to both geographical and time series visualizations.
curl -O https://raw.githubusercontent.com/tbaums/dcos-elastic-kibana-demo/master/expenditure-data.json
Run the below command to load the dataset in bulk to the expenditure
index and hospital
type in the Elastic cluster.
curl -s -H "Content-Type: application/x-ndjson" \
-XPOST http://coordinator.elastic.l4lb.thisdcos.directory:9200/expenditure/hospital/_bulk \
--data-binary "@expenditure-data.json"
Don't forget to logout of your SSH session
On your local machine, save the JSON blob below as kibana-config.json
to pass as options when we install Kibana.
{
"service": {
"name": "kibana"
},
"kibana": {
"elasticsearch_url": "http://coordinator.elastic.l4lb.thisdcos.directory:9200/",
"xpack_enabled": false
}
}
Then, run the below command to install Kibana on the cluster:
dcos package install kibana --yes --options=kibana-config.json