Hastic Graph Panel

A better version of Grafana's default Graph Panel. Able to render Anomalies & more.

Please note that we are still in alpha, so features are subject to change

Prerequisites

• hastic-server
• Grafana >= 5.1.1

Installation

• Navigate to /plugins directory in <GRAFANA_PATH>/data/plugins

• Download hastic graph panel

wget [PLACEHOLDER]

• Unpack downloaded files

tar -zxvf [PLACEHOLDER]

Installation from source

Note, that <GRAFANA_PATH>/data/plugins directory will NOT be present until grafana-server was started at least once!

Also possible to create directory manually by typing:

mkdir -p data/plugins

• Navigate to /plugins directory in <GRAFANA_PATH>/data/plugins

cd $GRAFANA_PATH/data/plugins

• Clone the repository

git clone https://github.com/hastic/hastic-grafana-graph-panel.git

• Navigate to /hastic-grafana-graph-panel directory

cd $GRAFANA_PATH/data/plugins/hastic-grafana-graph-panel

• Install necessary dependencies

npm install

• Build hastic graph panel

npm run build

• Restart grafana-server
  • For grafana installed via Standalone Linux Binaries:
    • Stop any running instances of grafana-server
    • Start grafana-server by: $GRAFANA_PATH/bin/grafana-server
  • For grafana installed via Package Manager:
    • type in systemctl restart grafana-server

Usage

hastic-server should be running in order to use anomaly detection.

• Open new dasboard where you want to see Hastic panel
• Open Dashboard Settings (top right corner) and then navigate to Variables
• Add new variable
  • set type to Constant
  • set name to HASTIC_SERVER_URL
  • set value to URL of your hastic-server instance in your dashboard (e.g. http://localhost:8000)
• Save settings and close Settings window
• Open panel edit mode (click panel titlemenu or press "e" hotkey)
• Navigate to Metrics tab. Set metrics. Only one metric suported
• Navigate to Analytics tab
  • Push Add an Anomaly Type button
  • Set name of the anomaly and choose a pattern type
  • Press create
• Label your data:
  • Click button with chart icon
  • Highlight anomalies on graph holding Ctrl button on Windows or Cmd on Mac
  • When you have finished labeling - click the button with the chart icon once more. saving... status should appear.
• Learning status should appear while hastic-server is learning (first learning can take a while).
• When Learning status dissapears - the anomalies should become labeled in your graph
• this icon means that the anomaly was marked by the server
• this icon means that the anomaly was marked by the user

Development

Build

npm install
npm run build

Changelog

[Improvements]

• You can zoom during update

Credits

Based on

• grafana-plugin-template-webpack-typescript
• @types/grafana