Limitations on Private Cloud
There are some limitations to keep in mind when you are working with Cloudera Machine Learning on Private Cloud.
The following features are not yet supported in CML Private Cloud:
- Logging is limited, and diagnostic bundles for each workspace cannot be downloaded from the workspace UI. Instead, diagnostic bundles for the entire cluster can be downloaded from the control plane.
- Monitoring on Private Cloud does not support node-level resource metrics, hence only K8s Cluster and K8s Container dashboards are available.
- ML Runtimes are not supported.
- CML does not support the NVIDIA Multi-Instance GPU (MIG) feature.