2.6 KiB

Raw Blame History

Hastic server

Implementation of basic pattern recognition for anomaly detection.

Implementation of analytics unit for Hastic.

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

Download & Install on Linux

You need only nodejs >=6.14 on your machine

wget https://github.com/hastic/hastic-server/releases/download/0.1.0-alpha/hastic-server-0.1.0-alpha.tar.gz
tar -zxvf hastic-server-0.1.0-alpha.tar.gz
cd hastic-server-0.1.0-alpha/server/dist
node server

Build & run from soure

Hastic server requires Grafana's API key (http://<your_grafana_url>/org/apikeys) to query data from Grafana datasources. API key role requires only Viewer access.

Possible to install on:

Linux
Docker

Linux

Environment variables

It is possible to export the following environment variables for hastic-server to use:

HASTIC_API_KEY - (required) API-key of your Grafana instance
HASTIC_PORT - (optional) port you want to run server on, default: 8000

System prerequisites:

git
nodejs >= 8.x, but there is special build for nodejs 6.14
python3 with pip3

Installation

pip3 install pandas seglearn scipy tsfresh

git clone https://github.com/hastic/hastic-server.git
cd ./hastic-server/server
npm install 
npm run build

Run

export HASTIC_API_KEY=<your_grafana_api_key>
export HASTIC_PORT=<port_you_want_to_run_server_on>

cd ./hastic-server/server
npm start

Docker

Build

git clone https://github.com/hastic/hastic-server.git
cd hastic-server
docker build -t hastic-server .

Run

docker run -d --name hastic-server -p 80:8000 -e HASTIC_API_KEY=<your_grafana_api_key> hastic-server

Known bugs & issues

Adding labeled segments while learning is in progress is not supported
Dataset doesn't update after 1st learning
Currently only influxDB datasource is supported

2.6 KiB Raw Blame History