(A better way to) Monitor Your Parity Substrate Node

I’ve spent the past few days learning more and more about Prometheus, Grafana, and how both of them work together to monitor services and clients running on your computer.

Specifically, I wanted to monitor my Ubuntu server which is running a couple of substrate validators (Kusama and Centrifuge). I discovered a few online guides and dashboards which were a bit outdated and didn’t deliver the useful metrics and experience I was looking for.

Here’s is what I learned and created.

1. If you haven’t already, install Prometheus and Grafana on your server. You can follow the simple instructions on this page. Stop when you get to the part where you import the Grafana 12425 dashboard.
2. Instead of importing the 12425 dashboard (which is cool, but outdated), import the 13697 dashboard for your Polkadot / Kusama validator or the 13698 dashboard for your Centrifuge validator. I created both of these by tweaking the 12425 dashboard created by phillux.

That’s it! You now have a sweet Grafana dashboard for monitoring your Polkadot / Kusama / Centrifuge validator(s).

Now, it’s time to install the Node Exporter software so you can monitor the core metrics of your server (i.e. CPU usage, memory usage, network usage, etc).

1. Navigate to Step 11 on this page for instructions on how to install Node Exporter and run it as a service.
2. After installation, import the very awesome 1860 dashboard created by rfraile.

And, you’re done! You now have a dashboard that displays key metrics of your server. Feel free to set alerts on any of the dashboard panels so you’re notified if your server is operating abnormally.

I hope you find this short tutorial and guide helpful.