Our previous monitoring tool tracked ping latency (server → target).
I have been asked to implement similar network latency tracking in Prometheus.
I don’t think it needs to be ping/ICMP specifically but I know that could be an option with the blackbox exporter.
We monitor a few hundred Windows servers via the windows exporter.
The exact milliseconds of latency is not all that important in this case. We are more concerned about trends. For example, if a service is suddenly much slower than normal, can we determine if the slowness is the result of high network latency or an issue with the application itself.
If I am trying to determine network latency, would the following query get me a result analogous to (i.e. similar to and roughly as useful as) a ping?
scrape_duration_seconds - windows_exporter_scrape_duration_seconds
Or are there other factors (e.g. responsiveness of http server) that this query takes into account that could muddy the results? Or should I just configure the ICMP monitoring in the blackbox exporter?
