Version 1.6.0+
These are the release notes for the v1.6.0 release of witboost.
Please refer to the official documentation for a more in-depth overview of the released features.
Features
New Notifications and Webhook channel
This release brings a new notification engine and new notifications that enhance the data developer and data consumer experience:
- System Fix Needed: A runtime policy detected a non-compliant system where you are either owner or part of the development team.
- System Operations Impact: A runtime policy detected a non-compliant system. Consumer systems operations may be affected by incoming upstream data.
- System Update: An existing data product is deployed and updated in the marketplace on one of the configured triggerEnvironments.
- New System: A new data product is deployed and published in the marketplace for the first time.
Moreover, all of the above notifications, including Access Control Responses can also be delivered through webhooks or emails.
Azure DevOps integration
Starting from this release, Witboost supports the integration with Azure DevOps. This integration allows you to:
- create new systems and components in witboost by storing them in Azure DevOps repositories
- define new templates, blueprints, domains, etc stored in Azure DevOps repositories
- include an action in your templates to create new Azure DevOps repositories for your entities
Added the editor page for components
The "Edit & Test" tab is now available for components belonging to a system. This tab allows you to:
- edit the descriptor of the component (directly on the repository or leveraging import/edit mechanisms)
- see the whole descriptor of the system
- test the complete descriptor
Since deployment is enabled only for systems, the release and deployments tab is not available for components. You can still deploy a component by deploying the system it belongs to, in its dedicated tab.
Partial provisioning per environment
Partial provisioning is a feature that enables the user to select subsets of the systems to deploy. This feature is useful when you want to deploy only a handful of components belonging to a system in a specific environment (e.g. for testing purposes).
Since this behavior breaks the system's "atomicity", it is disabled by default. To enable it, you need to add the following configuration to the configuration file:
mesh:
provisioner:
partialProvisioning:
enabled:
- development
- test
In the configuration, you can specify the environments where partial provisioning is enabled. When enabled for an environment, the user can select the components to deploy in the system's deployment page; by selecting the "Advanced Provisioning" option, the user can choose the target status for each component. For example, the user can decide to redeploy an already deployed component, undeploy an existing component, and skip the deployment of a new component, all in the same request.
User Interface revamp
The blueprint page has been revamped to improve the overall user experience. Now it should be easier for the user to understand what entities are being created and what steps are required to complete the overall flow.
Regarding the theme, we added the possibility to customize the colors of:
- the Data Product Graph
- the colors used by charts and graphs (e.g. Mesh Supervision, Governance Overview, Marketplace Dependencies, etc.)
Improved policies creation
Since policy and metric creation was a bit confusing and required a long time, we decided to improve the user experience by adding default values for the form that guides the user through the process of creation.
The default values are the most common ones, proposing the creation of a deployment active policy. The user can still change all the values if needed, but the form is now more user-friendly.
Deployment information propagation
Every time a system is deployed, all the technology adapters will propagate their deployment information to the technology adapters scheduled after them.
Let's picture a scenario where you have a system with a component B that depends on component A: for example, B could be a view on top of storage area A, or B could be a workload reading from table A, or any pair of components for which B needs a reference to A to be created. In this case, B could need some details after A is provisioned to be able to create the view or the workload (e.g. the table name, the storage area name, etc.).
With this feature, every technology adapter will receive the deployment information of all the previous technology adapters, so that the information can be used to create the next component.
Issues Fixed
- Added missing field on policy creation form when resource type is 'Data product' and _trigger_is 'active'
- Resolved the Unexpected end of JSON input error in the access control page
- The logs drawer does not become blank anymore when deployment finishes
- It is now impossible for the user to enter arbitrary values for the DescriptorPicker
- Handled default values for EntitySelectionPicker
- Fixed a bug that prevented the display of a metric result if the value was 0
- Fixed date filtering in CatalogTable when selecting only one day
- Pickers can fetch the information regardless of their position on the form
- Disabled auto escape when parsed string value with nunjucks to resolve some string visualization problems
- Descriptor's configurations from branches different from the master one are now fetched correctly
- Fixed the test/update or create a new release if the version in the descriptor is not the same as the one in the current data product
- Fixed the appearance when the structure panel would be very narrow when the test panel is open and long-named components might overflow
- In the release table page, the update button and the new draft release button now send the right branches for each component
- Added the shortcuts for deploy and undeploy in the release table
- Fixed crash if data sharing agreement field does not exist in the descriptor
- Fixed a bug that caused a no refresh on the microfrontend page when switching between microfrontends
- Fixed a bug that forced users to click out of the box to close the autocomplete suggestions in the IdentitiesPicker
- Enhanced validation process: Previously, during testing, if a component had its "specific" attribute set to null, it would be inadvertently skipped. This issue has been addressed, and such components are now properly included in the validation checks
- Fixed a bug where EditTemplate picked values from previous edits
- Fixed an inconsistency with the CustomUrlPicker
- Fixed markdown rendering in templates
- Fixed an error when requesting access to an Output Port from its page
- Fixed a bug that prevented CGP to generate a runtime policy failed event even for local policies
- Fixed a bug on CGP where a policy result was accepted even if the resource did not exist
- Fixed a bug that failed builder operations due to missing authentication
- Fixed a bug where the builder incorrectly retrieved the branch name of a resource
- It is now possible to choose between
Undeploy
andUndeploy and remove data
in advanced provisioning when undeploying a resource - Fixed a bug where the builder could show a
Environment not set
error when opening the Edit and Test tab - On an edit operation, if the skeleton of the edit template defines the field
witboost.parameters
it is now prioritized - It is now possible to have the character
:
in your password when using LDAP - Fixed a focus problem in the edit panel
- Fixed a crash in the test panel
- Fixed a bug where the scheduler did not correctly handle waiting tasks for all dependencies