On Monday, 29 June 2020 18:27:07 UTC+1, Kevin Vasko wrote: > > ahhhh, > > Okay. That’s it. I was using the gpu node exporter for those machines and > apparently it doesn’t have the node_uname_info variable. > > Just curious, what is the best practices for when to use different > “job”(s)? > > I broke it down into “cpu-nodes” and “gpu-nodes” based on the exporter. > I’m getting the feeling this isn’t the best idea. > > No, that's absolutely right and proper - different jobs to scrape different exporters.
What you probably want to do is to control the instance labels, so that the metrics from both jobs both use instance="machine", rather than instance="machine:9100" and instance="machine:9200" (say). https://www.robustperception.io/controlling-the-instance-label This can be done with a little bit of label rewriting in your scrape jobs. Then you can make queries which join between node_uname_info and nvidia_gpu_power_usage, using the common instance label. -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/8b6c6285-3bb4-499c-9348-1f0646216f6fo%40googlegroups.com.
