What we're seeing is a data lag between Jenkins and Rundeck, where Jenkins is tailing the Rundeck job output (it's a very long running job with lots of output) and after 2.5 hours, Rundeck says it's completed successfully, but Jenkins is still tailing an active log with output. It's almost like a flush hasn't happened between Rundeck and Jenkins.

Eventually Jenkins logs will show the same results as Rundeck, but it's not synchronous with Rundeck's actual output.

This is not reported when rundeck jobs has a small output, so it happens with large data on the rundeck job's output.