grafana-stack

Overview

The goal of this directory is to set-up the following monitoring stack:

• InfluxDB as the Time-series Database
• Grafana as the front-end/display
• Prometheus as a different flavor of time-series database
• Nginx is used as a proxy (mostly to fix CORS issue)

Testing locally

You can set-up your own local grafana-stack easily, it doesn't have the same constraints as production: No need to set-up a specific password or to go through the nginx proxy.

You can run:

docker run -d -p 8083:8083 -p 8086:8086 tutum/influxdb:0.13
docker run -i -p 3000:3000 grafana/grafana
docker run -p 9090:9090 -v prom/prometheus

You then need to set-up the datasources/users into Grafana and InfluxDB as described below, and then import dashboards from dashboards/ directly through the web interface.

Step-by-step

First-time only

Create the InfluxDB password:

kubectl create secret generic influxdb --from-literal=rootpassword="${influxdb_password}"

Password and database creation

Connect to the container and create the db:

$ kubectl exec -i -t influx-${PROJECT}-123456 -- influx -username root -password "${influxdb_password}"
> create database github

Adding data-source

Probably a first time only:

./datasource.sh ${nginx_ip} ${grafana_password}