Infrastucture for Prometheus & Grafana monitoring

Install Node_exporter in the servers you want to monitor. Install Prometheus & grafana in the central server that will monitor all the nodes.

---

Node_exporter

This will create a server in 0.0.0.0:9100/metrics with all the metrics of the machine. Then, you pass the_ip_of_the_machine:9100 to prometheus to make it gather the data.

To create the service, run:

docker-compose -f docker-compose.node_exporter.yml up

To test if the service is working, run:

curl localhost:9100

To destroy the service, run:

docker-compose -f docker-compose.node_exporter.yml down

---

Prometheus

This will create a service in 0.0.0.0:9090 to gather all the data from the target node_exporters

To create the service, run:

docker-compose -d -f docker-compose.prometheus.yml up

To destroy the service, run:

docker-compose -f docker-compose.prometheus.yml down

Edit the prometheus.yml file in the config folder to add ips of target machines with node_exporter running

static_configs:
  - targets: ['server1_ip:9100','server2_ip:9100']
    labels:
      groups: 'my_servers'

---

Grafana

This will create a web graphical user interface to import and create dashboards for prometheus data

docker-compose -d -f docker-compose.graphana.yml up

To destroy the service, run:

docker-compose -f docker-compose.graphana.yml down

open a browser and go to:

• ip_of_this_machine:3000
• user: admin
• password: admin
• create new dashboard
  • configuration >> data sources >> add data source >> prometheus >> url: http://localhost:9090 >> save & test
  • create >> import >> 1860 >> load >> <my_dashboard_name> >> select prometheus >> create