CloudBees CI on modern cloud platforms 2.277.4.2

In the operations center, CasC bundles can now be assigned to more than one controller. Additionally, the administrator can globally change or modify the default bundle behavior, allowing a single configuration bundle to be used by all controllers. Once a bundle has been verified with one controller, it can be used across one or multiple controllers. This eliminates the need to create a new bundle each time you set up a new controller thus reducing time, manual error, and overhead. Assigning a bundle to more than one controller is a Preview feature.

The CloudBees Software Delivery Automation navigation bar has been updated to include a new Help menu that contains useful links for CloudBees documentation, release notes, onboarding, troubleshooting, support, and providing product feedback.

To allow a configuration bundle to be assigned to one or more controllers, the assignment of a bundle to a controller has been moved from the Configuration as Code bundles screen to the controller configuration screen in operations center.

When the nectar-rbac plugin fails to read its configuration at startup, it uses the default authorization, granting administrative permissions to all authenticated users.

When you attempt to deploy CloudBees CI on modern cloud platforms on EKS with Kubernetes 1.19 or higher using the Helm chart, the deployment fails with an error on an unknown field service name.

When the Kubernetes MinorVersion value contained a +, the Helm chart chose the incorrect Ingress. This issue was occurred mainly in cloud environments (EKS/GKE/AKS) and Kubernetes forks.

When a controller falls back to the offline security realm, some security realms such as LDAP fail. The configuration is incorrect due to a serialization issue.

When you define a shared configuration in operations center, and you add a mail configuration snippet, the validation fails. The validation method is broken due to a compatibility problem with the mailer plugin.

The URL used to download pip for Python was broken, and a manual fix was required to install the AWS CLI.

If only pip for Python 3 is installed on a system, the python-pip plugin is unable to install the AWS CLI.

A missing permission check allowed users with read permissions to reload custom update centers.

An authentication call was not being activated when users logged in, causing audit implementations to miss it.

On EKS, the ALB Ingress Controller annotation to define the health check path was not defined, causing the target groups to be unhealthy.

Previously, users could access the move/copy/promote logs without the proper permissions.

Before this change, HTTP communications were handled by an old version of async-http-client that didn’t support Server Name Indication (SNI).

CloudBees is updating terminology to remove offensive text. During this ongoing initiative, “controller” replaces “master,” “agent” replaces “slave,” “allowlist” replaces “whitelist,” and “denylist” replaces “blacklist.”