This change adds mappings to the graphite exporter configuration file that map the k3s metrics into unified prometheus metrics. When running apps in TrueNAS scale the pods are exported by the metrics exporter and can be captured here to get some basic operational details.
❗ The metrics exported are not documented in TrueNAS or k3s in this format, and this poses a problem making sense of the cpu metric.
I've been able to identify that these numbers go up/down in the 100,000 -> 10,000,000 range for a mostly idle system. I suspect they're something like CPU time in... nanoseconds? over some period. I will admit I'm not an expert at system metrics to make a better educated guess.
This change adds mappings to the graphite exporter configuration file that map the k3s metrics into unified prometheus metrics. When running apps in TrueNAS scale the pods are exported by the metrics exporter and can be captured here to get some basic operational details.
❗ The metrics exported are not documented in TrueNAS or k3s in this format, and this poses a problem making sense of the
cpu
metric.Example Metrics: