Using Netdata for Desktop Stats

An Observation

While the native task manager and htop utility that are shipped with most desktop Linux distributions are adequate for a general overview of system resources, they don’t provide a great UI for delving into the deeper metrics.

It would be really neat to have an ultra-detailed overview of system metrics across CPU/RAM/Disk, networking, cgroups, applications proceses, etc.

A Suggestion

Primarily usd for servers and other remote systems, Netdata is the de facto monitoring tool for collecting and visualizing system usage data. Netdata captures everything, no really it does.

While the installation of Netdata is geared towards servers, it’s pretty straightforward to install it on your Linux desktop to achieve the same effect. So why reinvent the wheel?

Release the Kraken

Open up a terminal, and run the following command:

bash <(curl -Ss https://my-netdata.io/kickstart.sh)

That will build Netdata from source (probably asking for sudo permissions mid-way) based on your Linux distribution, and install the correct system services and a monitoring dashboard that you can immediately access via http://localhost:19999.

Put a Leash on It

You might want to head over to Netdata’s Security Overview in order to set sensible defaults. I would recommend just ensuring that the web interface is accessible only from localhost using the following edits to the /etc/netdata/netdata.conf file:

[web]
allow connections from = localhost

Further Thoughts

Now that you can get pretty sweet performance insights across every aspect of your system, you may want to consider the following:

Run a local instance of Prometheus for a time-series database for long-term storage of your system metrics
Install Grafana to query said Prometheus TSDB just because you can
Customize the metrics that Netdata collects to help match those of your development environment or custom set of applications

04 Apr 2020 guides