1.45.0
Release notes
✨ New features
Databricks Unity Catalog
Deploy your model from the Databricks Unity Catalog using the Databricks integration! Configure the Databricks credentials on the ingrations page and make use of a reference.json
to point to the desired model to deploy it.
Temporary credentials
Use temporary credentials to safely authenticate to your object storage. Do this by filling in AWS S3 temporary credentials. Azure Blob Storage's SAS token or a Databricks temporary access token in the Deployment flow.
Dutch algorithm register integration
Connect your Deployment to the Dutch government's Algorithm Register. Add algorithms via your Deployment compliance or link your Deployment to an existing algorithm to get started!
🛠️ Improvements
- Add support for MLFlow model version alias
- Improved performance of the deployment and prediction log overview pages
🐛 Bug fixes
- Fixed an issue where failure in deployment artifact download was not reflected in the event logs
- Fixed an issue where teams could not be deleted (on Deeploy cloud)
- Fixed an issue where Workspace with assigned credentials could still be selected in the integration Workspace permissions
- Fixed an issue where the activity graph could disappear in the deployments overview
- Fixed an issue where the activity and response time graphs showed incorrect data with certain filters
- Fixed an issue with selecting compliance documentation templates in the create Deployment flow