Traditionally, PiCluster had the ability to store its log output in Elasticsearch. Starting in PiCluster 2.3, node metrics such as disk, cpu, and memory utilization is stored also. With this ability, users can easily view their host metrics over time.
Requirements
- PiCluster
- Elasticsearch 6+
- Kibana 6+
If you want to quickly spin up an Elasticsearch cluster with Kibana in Docker, check out my elasticsearch-demo in my Git page.
PiCluster Configuration
PiCluster needs to be configured for Elasticsearch so it knows where to send data. All you need to do is add the following config option in config.json that points to your Elasticsearch endpoint and restart the PiCluster server process.
"elasticsearch": "http://my-es-endpoint:9200"
Kibana Configuration
Starting in PiCluster 2.3, there are pre-set graphs available for viewing host metrics (cpu, memory, and disk).
Creating the Index Pattern
- Click on
Management->Index Patterns. - Click
Create Index. - Under
Index Pattern, typepicluster-monitoring - Click
Next Step - Under
Time Filter field name, choosedate. - Click
Create. Repeat steps 1-6 for thepicluster-loggingIndex Pattern.
Importing the Graphs
To import the PiCluster Monitoring graphs in Kibana:
Click on Management -> Saved Objects.
Click import.
Navigate to the PiCluster Git directory and choose kibana-graphs.json
Click open.
Repeat steps 1-4 for kibana-dashboard.json