2.0 KiB

Raw Blame History

Hastic server

Implementation of basic pattern recognition and unsupervised learning for anomaly detection.

Implementation of analytic unit for Hastic.

Build & run

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

Docker

Example of running hastic-server in 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

Linux

Environment variables

You can export 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

Dependencies

You need in your system:

nodejs >= 6.0.0
python3 with pip3
curl gnupg git make g++

Intallation

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

Known bugs & issues

If you add labeled segments while learning - it fails
Dataset doesn't get updated after 1st learning
Currently only influxDB datasource is supported

2.0 KiB Raw Blame History