Creating databases/graphs from motd information using RRDtool

bomberb17 · 11-24-2022, 08:55 PM

Hello,

when I log to my system (Ubuntu Server 20.04) I get the motd as follows:

Code:

  System information as of Fri Nov 25 02:48:55 UTC 2022

  System load:           2.01
  Usage of /:            64.7% of 1.79TB
  Memory usage:          94%
  Swap usage:            5%
  Temperature:           47.7 C
  Processes:             147

I would like to use RRDtool to save all these numbers (i.e. the system load, disk usage, memory usage (with swap), temperature and processes) and use it later to create nice graphs.
Is there a guide/tutorial on how to do this?
The closest tutorial I could find is this one
https://calomel.org/rrdtool.html
which however I don't know how to modify to get the numbers from motd.

elgrandeperro · 11-25-2022, 10:28 AM

the dynamic motd runs scripts in /etc/update-motd.d. So you can get the script there that spits out those values in a file then you can graph them.
If you want it as an exercise or learning, fine.

Personally, I'd just install nagios and gets some existing plugins to localhost. You will get a way to display them, easily configured and graphed.

For instance, for system processes:

https://exchange.nagios.org/director...rics/Processes

bomberb17 · 11-25-2022, 11:08 AM

The host is a Raspberry Pi which runs close to its limits, and I would rather make the log as lightweight as possible.

So I see the motd information output is from

Code:

/usr/bin/landscape-sysinfo

. Maybe instead I could just take the numbers separately, i.e.

Code:

uptime

,

Code:

df -h /dev/sda1

etc.
But my question is still on how to use these for updating a rrd database, i.e. in the example I posted, how would the

Code:

update_rrd_database.sh

script be for monitoring the system load?