I collecting from multiple hosts metrics via Telegraf and sending them to an single DMZ Telegraf (inputs.http_listener) instance. This DMZ instance sends all metrics to a InfluxDB, where then I can view them. Now I need the same data also on a Prometheus installation (Thanos), so I enabled the outputs.prometheus_client on the DMZ instance, where I can then get the data also from Prometheus, but only from the instance itself (the DMZ node) , not from all others, which using this one as relay.
Is there are any special config required in Telegraf ?
The filter … that would mean, the “instance” should replaced with the original hostname ?. At the moment … I can see the data from all nodes inside the DMZ node. It means: On Grafana I have all nodes in a single instance … which brings your laptop fan to new levels of noise.
I’ve changed my Dashboard for testing and using the host label … to split them, instead of instance. But I have a different problem … I have a lot of holes. I will start a new thread for this problem.
