monitoring-grafana-influxdb.../README.md

2.1 KiB

What is in this repository?

This repository contains easily deployable monitoring solution which uses:

  • Grafana (frontend for monitoring + alerts)
  • Prometheus (monitoring solution pulling metrics from exporter)
  • Node Exporter for Prometheus (metrics exporter-exposer for Prometheus)
  • Telegraf (monitoring agent)
  • InfluxDB (persistent timeseries storage)
  • cAdvisor (containers monitoring)
  • alertmanager (alerting)

How to use it?

If you have docker and docker-compose installed, this will take roughly 1 minute to have it up and running. If not - it will still take mentioned ~ 1 minute + time needed for docker installation.

Here is how to install:

With certificates and HTTPS (optional):

Prepare certificate in /etc/grafana/ (privkey1.pem, fullchain1.pem).

Uncomment following lines in docker-compose.yml:

...
      volumes:
        - grafana_lib:/var/lib/grafana
#        - /etc/grafana/privkey1.pem:/etc/grafana/privkey1.pem:ro
#        - /etc/grafana/fullchain1.pem:/etc/grafana/fullchain1.pem:ro
...
       environment:
         GF_AUTH_ANONYMOUS_ENABLED: "false"
         GF_SECURITY_ADMIN_PASSWORD: "nimda321"
...
#        GF_SERVER_CERT_FILE: "/etc/grafana/fullchain1.pem"
#        GF_SERVER_CERT_KEY: "/etc/grafana/privkey1.pem"
...

Also uncomment these lines in "deploy_all.sh":

## NOW LET'S SECURE GRAFANA
# CHECKING OUT ORIGINAL FILE
echo -e "checking out original docker-compose.yml"
git checkout docker-compose.yml

## STOPPING and REMOVING GRAFANA CONTAINER
echo -e "stopping & removing grafana container"
container_id=$(docker container ls | grep grafana| awk '{print $1}')
docker stop $container_id
docker rm $container_id

# REPLACING HTTP with HTTPS
echo -e "changing http to https"
sed -i 's/GF_SERVER_PROTOCOL: "http"/GF_SERVER_PROTOCOL: "https"/g' docker-compose.yml
docker-compose up -d grafana
echo -e "reverting: changing https to http"
sed -i 's/GF_SERVER_PROTOCOL: "https"/GF_SERVER_PROTOCOL: "http"/g' docker-compose.yml

Deployment:

$ ./deploy_all.sh

Monitoring should be up and running http://_**hostname**_:3000/ or https://_**hostname**_:3000/