N.U.T - UPS Monitoring via PfSense / Grafana

N.U.T - UPS Monitoring via PfSense / Grafana

Server owners love electricity, that's why I'll tell you how I monitor my UPS to prevent my servers from shutting down the hard way.
N.U.T (Network UPS Tools) is a good choice to unify multiple brands UPS for monitoring.

I didn't want my hypervisor to assume this role, so I made some research and I saw a PfSense package for N.U.T.

My UPS is a small EATON UPS - MODEL ECO 800.
EATON---ECO-800-1

And it has an USB type B port on the back.
I connect the UPS to my PfSense box (WatchGUARD XTM 525, running PfSense)
PfSense

PfSense

Package installation

Go to the package manager and search for the nut package

1
Install it
2
3-1

Configuration

The configuration is quite easy, you just need to go to the settings page of N.U.T, and choose the right driver.

4

To know which driver you need, the N.U.T team did a hardware compatibility list : https://networkupstools.org/stable-hcl.html
Here's the research for my model
5-5

And ... We give a name to the UPS (placard in my case) configure the correct driver, we define the UPS type as USB
5

After this, we get back to the UPS status page to see the result
6

The_Greatest_Showman_TADA-1

Deeper exploitation

Now, our Firewall know the charge of the UPS, that's pretty nice, but, we could make something more personnal

The installation of N.U.T come with a command line tool called upsc, it displays all the information about your UPS.

upsc [ups_name]@ip_of_the_nut_server

9-5-1

Now, we format the data for exploitation
7

Grafana integration

For this part, I assume that you have already did a configuration of telegraf in PfSense (if not, refer to my previous guide : https://blog.lbdg.me/pfsense-monitoring/ )

I write a very simple script

filename : ups.sh, localized in /usr/local/etc
#!/bin/sh

charge=$(upsc placard@127.0.0.1|grep 'battery.charge:' | cut -d ' ' -f 2)
load=$(upsc placard@127.0.0.1| grep 'ups.load:' | cut -d ' ' -f2)

echo "ups,ups_name=placard,host=10.0.0.1 charge=$charge,load=$load"

Here's the script in action :
8-1

I unplugged the UPS to show you the charge decrease

I create an input filter in telegraf

Configuration add to /usr/local/etc/telegraf.conf
[[inputs.exec]]

        interval = "2s"
        commands = [
                "/usr/local/etc/ups.sh"
        ]

        data_format = "influx"

I reload the Telegraf agent and ... Data are coming in InfluxDB, here's a visual representation of the data with Grafana
10

Zabbix integration

I also have a Zabbix server and I wanted to send a command when the charge of the UPS is lower than a certain percentage.

I use remote commands to extract the data I need
11

I create 3 items
12

Here's the result
13

I create a trigger, if the ups.status is different than OL (On Line) and ... You can see a problem "battery connexion"
14

And a mail notification
15