Monitoring Linux Processes Using Prometheus and Grafana

Using bash and push gateway to go against the pull design of Prometheus? Why not just use Telegraf (daemon in go that use few resources) that already have a push design, to send the metrics to InfluxDB and then do the dashboards in Grafana? Or the same setup, but with Telegraf emulating being a Prometheus exporter.

Besides that, Telegraf have builtin how to gather a lot of metrics, that can give a better insight of what is happening in your server and all its components. And Grafana have ready to download a lot of nice dashboards to display that information.

Also, if you want to practice how to send your own and very particular metrics with shell script, InfluxDB's line protocol is very simple and you can just use curl to send them with shell scripts.