CloudBees CI on traditional platforms 2.516.2.28983

The following plugins have been added to the operations center CloudBees Assurance Program since 2.516.1.28669:

The following plugins have been added to the controller CloudBees Assurance Program since 2.516.1.28669:

After the switch to HTTP-based transport for shared agent/cloud provisioning in version 2.440.1.3, operations center printed log messages from a forked process of an inbound agent connecting to a client controller. However, this lacked adequate context to determine which client controller or agent was involved, particularly when many provisioning requests arrived in short order. Now, every such log line is prefixed with the name of the client controller and the label expression being provisioned.

Resolved an issue where orphaned EC2 instances could accumulate in the Amazon Web Services (AWS) console when using the Amazon EC2 plugin (ec2) configured to avoid using orphaned nodes. This update introduces a periodic cleanup mechanism for EC2 nodes that are no longer attached to active controllers. This mechanism prevents the unintentional buildup of EC2 instances, which contributes to resource exhaustion and breaches of plugin global caps.

Backups to Amazon Web Services S3 failed immediately with the message Your proposed upload exceeds the maximum allowed size when the backup size was greater than 5 GB. This was caused by a recent migration to the Amazon Web Services SDK v2, during which multipart upload was not preserved. The upload now uses a multipart upload and can handle large files.

Under certain heavy load circumstances, a High Availability (HA) controller may have allowed multiple replicas to adopt the same build in close succession, leading to corrupted build metadata.

Resolved a race condition that could occur when an SSH agent was simultaneously connected to multiple High Availability (HA) replicas. This issue was more likely to occur in environments with a large number of SSH agents and high parallel build activity. In such cases, multiple replicas could mistakenly schedule and run two builds concurrently on the same agent.

Offline cloud agents were previously displayed only in their owning replica. This led to a confusing user experience. Now, these agents are displayed in every replica.

If a Pipeline build was using an agent that gets disconnected and then stays disconnected for five minutes, it will now abort the sh step automatically. As usual, the retry step with conditions can be used to restart the stage if this was a temporary issue.

Backups to Amazon Web Services S3 could fail immediately when uploading the generated backup file. This occurred particularly when using AWS credentials with an IAM role. The client used to upload the backup was initiated at the beginning of the backup operation. Therefore, the issue occurred when generating the backup file took longer than the configured Security Token Service (STS) Token Expiration period. The session token used for uploading is now initiated right before the upload.

Using a file-protocol URL in a plugins.yaml file would typically result in a FileNotFoundException error in the system log. This behavior could have lead to confusion or misconfiguration during plugin management.

When you search for a specific plugin under the Available tab in the Operations center Plugin Manager, the search results show duplicate entries for the plugin.