duritong
duritong
The way through https://github.com/operator-framework/operator-lifecycle-manager/blob/master/doc/design/subscription-config.md#resources has the drawback that it configures this for *each* container. Meaning if you configure 1 CPU as a request, it actually requires 2, since kube-rbac-proxy will...
`* 8 * * *` would be every minute after 8 o'clock for one hour. If you want to start it at minute 0 you also need to set minute...
On nodes with H100 or H200 we are missing the DCGM_FI_PROF_GR_ENGINE_ACTIVE metric. Environment: OpenShift 4.16 / NVIDIA GPU Operator 24.9.2 The metric DCGM_FI_PROF_GR_ENGINE_ACTIVE is not reported at all on H200...