How to best monitor a bunch of raspberry pis running balena OS?

I am planning to run balenaOS on a bunch of raspberry pis running a multi container application (not all raspberry pis will run the same application).

I would like to monitor (e.g. in charts) the memory, cpu and disk utilization of those raspberry pis and the services running on them. The goal is to figure out if the raspberry pi is not yet overloaded and if it still has sufficient free resources to run more services ?

What is a good approach for doing this kind of monitoring ?

I was thinking of telegraf, influxdb and grafana.

thanks for any advice
Jan.

I definitely agree with your thinking on Telegraf/InfluxDB and Grafana - these are really awesome pieces of software. Depending on your situation you could consider setting up a central InfluxDB & Grafana instance and reporting all the statistics back to a central location rather than holding them on each device.

If you’re able to share, we’d be interested to see your final setup when you get it up and running.

1 Like