I need help to understand how can I get grafana deployed on the servers to get the metrics from our locally installed prometheus

grafana is deployed on (servers) kubernetes clusters. Prometheus is installed on rpi. I am not sure how to create a connection between them.

are they on same network, if not can they talk to each other?

There are not directly on the same network. I am thinking if I could save prometheus results to the server which both prometheus and grafana can access. Really not sure in what format to save the prometheus results. How to save it in the first place.

how about this question.

nope, the machine where I have the prometheus installed cannot communicate with grafana directly. It does via middle server which is a vm.

1 Like

and what does this vm do, what is it’s function? is it a proxy server? the less mysterious you are the more we can help unless you cant share due working with DOD or some gov agency

vm is nothing more than linux machine. Its an EC2 instance. Currently I am trying to get my RPI health metric to display on the grafana. The issue is grafana has restriction and can only interact with machines on its network.
I plan to save the prometheus result to data format like csv, mysql.
Push the data to the VM instance, and from their I want to use grafana to grab the data and display the metric.

1 Like

ok, now it is clearer. so vm can talk to both prom and grafa but prom and grafa cannot talk.

what you plan to do is good, but the csv approach would be doable but tedious for an automated continuous process.

The db approach is also good. but you can also bypass and cut out prometheus and use influxdb on the vm.

Why do you need prometheus for and then duplicate the same data in another vm if only grafana uses the prom data. Does that make sense? otherwise you will now create yet another process that will need to be synched with prom. also another option is to use node-red on your rpi to push data to the vm either mysql or influxdb