Closed horazont closed 4 years ago
This doesn’t work. The go-ping monitor/target only keeps history-size
amount of pings, even for the simple counters. Hence, this will never be properly accurate.
The only way out of this is a rewrite without the go-ping monitor things and instead using proper histograms.
This is a more natural way to represent the data in Prometheus. It also allows the user to create derived information such as loss rate over time, which is not possible with the loss_percent metric.
Fixes #24.