Open josericardomcastro opened 3 years ago
Can you please provide details of your platform? For example:
Prior to dcgm-exporter 2.1.2, we used to load profiling metrics by default and these metrics are only available on datacenter GPUs (i.e. 'Tesla'). Starting with 2.1.2, we added better error handling to prevent these errors and skip gathering metrics that are not supported.
My dcgm-exporter pod is crashing. Any ideias?