If a jobA triggers two other jobs (jobB and jobC), this would show as a split in a delivery pipeline view. If I want jobD to run after jobB and jobC completed (merge in the flow), I'd need to configure the Join plugin on jobA and say "run jobD once all downstream jobs have completed".

This works fine, but if say jobC failed and I rerun that stage in the delivery pipeline view, jobD doesn't get enabled. After testing it seems that whether jobD runs or not is determined by the initial assessment of the Join plugin. No matter what I rerun from the downstream jobs (jobB, jobC), jobD stays in its original state.