Major Enhancements
Restricting the total number of the 'Running' workspaces on a cluster
With this release, it is possible to restrict the total number of the 'Running' workspaces on a cluster using the maxNumberOfRunningWorkspacesPerCluster CheCluster CR property. More details about this feature can be found in the official documentation and the demo video.
Specifying the list of the allowed sources based on which CDEs can be started
It is now possible to specify the list of URLs based on which Cloud Development Environments (CDEs) can be initialized using the dedicated optional allowedSources CheCluster CR property:
"devEnvironments": {
"allowedSources": {
"urls": ["url_1", "url_2"]
}
When using this property, If a particular URL is not allowed explicitly by the admin, users will not be able to initialize and create CDEs based on this source:
More details about this feature can be found in the official documentation.
IntelliJ 'WebStorm' and 'PyCharm' editors for the remote development using JetBrains Gateway
IntelliJ 'WebStorm' and 'PyCharm' editors for the remote development using JetBrains Gateway are now available on the User Dashboard:
Passphrase configuration for SSH keys
Now it is possible to specify a passphrase while adding a new SSH key using the "SSH Keys" tab in the User Preferences:
Adding an option to deploy operands on specific cluster nodes
With this release, it is possible to deploy operands managed by the operator (dashboard, gateway, plugin-registry etc.) on the specific cluster nodes using the dedicated nodeSelector and tolerations properties CR properties:
dashboard:
deployment:
nodeSelector:
tolerations:
Possibility to hide some of the default editors on the User Dashboard
It is now possible to conceal editor definitions which is useful when an admin wants to hide particular editor(s) from the Dashboard UI, e.g. remove the IntelliJ and have only Visual Studio Code - Open Source visible. More details about the procedure can be found in the official documentation.
Allow starting existing workspaces even if GitHub is down
Before this release, if GitHub OAuth was configured and GitHub was down, a user could not start existing CDEs from this source. The defect has been fixed in this release and existing CDE will be started even if the OAuth provider is down with the dedicated warning message shown during the startup:
Support devfile endpoint annotations
With this release, endpoint annotations can be provided in the devfile. For example, the following devfile snippet will create an ingress or route with the annotation foo: bar on CDE startup:
components:
- container:
endpoints:
- name: my-endpoint
annotation:
foo: bar
...Provide a way to set the runtime class for all CDE pods using CheCluster CR
The spec.devEnvironments.runtimeClassName property has been added to the CheCluster CR. This property sets the spec.runtimeClassName for all CDE pods. The controller.devfile.io/runtime-class attribute for the CDE has precedence over the CheCluster spec.devEnvironments.runtimeClassName property.
Ignore the 'FailedScheduling' event by default when starting a CDE
With the release, the FailedScheduling event is the default value for the spec.devEnvironments.ignoredUnrecoverableEvents property. This is beneficial when the cluster has an autoscaler configured. If a CDE pod cannot be scheduled on any node causing a FailedScheduling event, the CDE startup will be retried when the new node is provisioned.
Bug Fixes
Dashboard is not available when using the CHE_FORCE_REFRESH_PERSONAL_ACCESS_TOKEN property
Previously there was a known issue affecting workspaces using Azure DevOps, Bitbucket or GitHub providers in connection with the CHE_FORCE_REFRESH_PERSONAL_ACCESS_TOKEN property. Every time you started a workspace, a new personal access token (PAT) was added and the previous one was not removed. When there were more than five existing PATs, you were not able to run the workspace, and the Dashboard was not available. The defect has been fixed in this release.
Do not show the previous CDEs error during restart
Before this release, the error message from the previous startup could have been shown after the CDE restart which was confusing for the user:
The defect has been fixed in this release.