Hi @BugRoger
From the Readme, we can see that you deploy the nvidia-exporter as a Deployment, it may not reasonable if I have more GPU worker nodes in my cluster. Should this be updated as Daemonset ?
Another question is : actually I find your work from grafana dashaboard, and want I would like to do is to expose nvidia metrics to grafana via kube-prometheus monitoring system.
However, seems I can NOT find your metrics in Prometheus service discovery after deploying your service. Seems I have to deploy service account and service monitor for it. Do you have any guidance or suggestions?
Cherishty
commented
5 years ago
Hi @BugRoger From the Readme, we can see that you deploy the nvidia-exporter as a Deployment, it may not reasonable if I have more GPU worker nodes in my cluster. Should this be updated as Daemonset ?
Another question is : actually I find your work from grafana dashaboard, and want I would like to do is to expose nvidia metrics to grafana via kube-prometheus monitoring system. However, seems I can NOT find your metrics in Prometheus service discovery after deploying your service. Seems I have to deploy service account and service monitor for it. Do you have any guidance or suggestions?
Best Regards!