What's New
Major features and updates plus known issues for Cloudera Data Science Workbench 1.10.0.
New features
- Cloudera ML Runtimes are in GA and the default and recommended solution to run user workloads. New projects will be created with ML Runtimes configured by default and we recommend migrating existing projects to use ML Runtimes. Legacy Engines are deprecated and will be removed in a future release but workloads running on them remain fully supported.
- New ML Runtimes 2021.09 are available including:
- NVIDIA GPU edition
- RAPIDS edition
- Scala and R support kernels
- New capabilities on ML Runtimes:
- Register customized Runtime - Administrators can register an externally built Runtime to provide Data Scientists with a customized environment.
- Runtime Catalog
- Connecting to secure Docker registries - Administrators can now add basic Docker credentials for external docker registries.
- CDSWCTL support for ML Runtimes
- Applied Machine Prototypes (AMPs) - Applied ML Prototypes (AMPs) are now generally
available inside the product, which provide end-to-end prototypes to help kickstart real
customer use cases. AMPs effectively package pre-built models for data scientists to tailor
for their own needs, as well as enable data scientists to learn how the different parts of
CML work together.
- APIv2 - A new API for operations on Projects, Jobs, Models, and Applications is now generally available.
- Business user experience - A new user role, ML BusinessUser, provides restricted access to view Applications created in CML.
- User and workload level environment variables now override global settings
- Support for Kubernetes 1.19.15
This release fixed various bugs that are below.
Fixed issues
- Applications
- Admin needs a way to see all applications.
Cloudera Bug: DSE-10231
- Admin needs a way to see all applications.
- Storage
- You can now delete Projects through the UI without leaving residual files on disk.
Cloudera Bug: DSE-393
- You can now delete Projects through the UI without leaving residual files on disk.
- Jobs
- Support the full capability of the CRON scheduler in Jobs
Cloudera Bug: DSE-6256
- Include project name in job status email subject line
Cloudera Bug: DSE-14259
- Job details history page showing detailed timestamps
Cloudera Bug: DSE-14867
- Engine image used in Jobs can be configured
Cloudera Bug: DSE-7426
- Support the full capability of the CRON scheduler in Jobs
- ML Runtimes/Engines
- Runtime selector provides initial values using python 3.7 kernels, not empty
options
Cloudera Bug: DSE-13769
- Runtime selector provides initial values using python 3.7 kernels, not empty
options
- MLOps
- Improved key technical metrics visualizations
Cloudera Bug: DSE-13742
- Improved key technical metrics visualizations
- Security
- Allow users to rotate API keys
Cloudera Bug: DSE-15715
- Switch to SHA256 for SAML signature validations
Cloudera Bug: DSE-16762
- Allow users to rotate API keys
- Performance
- Model build performance improvement
- Monitoring improvements
- Adding metrics on the internal database
Cloudera Bug: DSE-9129
- Adding metrics on the internal database
- K8s upgrade to 1.19.15
-
Third-party editors
- Sessions with Browser IDEs running now adhere to the limit set in
IDLE_MAXIMUM_MINUTES.Cloudera Bug: DSE-6651
- Sessions with Browser IDEs running now adhere to the limit set in
For a list of supported CDSW versions, see Support Lifecycle Policy
