I think it is convenient to collect and investigate the application status if go-application outputs metrics of the process.
For example, lora-gateway-bridge only listens on UDP protocol, so it is difficult to check the application health. In this situation, if process metrics is exposed, there will be no problems.
And, collecting the application metrics is necessary to research and trace the system performance.
Have you had a look at https://prometheus.io? They provide a client too for exposing metrics. The big win is that Prometheus integrates with Grafana, so you can directly graph all the metrics and set alerts