Recent changes to Jenkins seem to have made startup times longer, and have acerbated this issue, breaking several automations.  However— this issue hopes to address a deeper problem with the startup sequence.

The Jenkins initialization reports as completing, then other processes begin to make Jenkins "ready to work...".  It's completely okay that these processes will take a while and that the Jenkins application is in-place before it's able to respond to Web service events, say.  The problem comes when automations attempt to restart Jenkins, and presume that Jenkins has been restarted when the service reports as restarting.

At a deep level, the Jenkins context reports initialization as complete before functionality is available.  For the GUI user, we present a nice interface that clearly tells the user to wait.  However, for anything relying on Jenkins, there is no cue differentiating Jenkins being "up" and "waiting to come up."  Automations can wait for the service to start, but service wrapping report Jenkins as being started— in some cases many minutes before Jenkins is prepared to respond to any events.

This issue proposes that this is a bug, and that the init() process should wait for dependent processes, and not return as complete until Jenkins is considered fit-for-use.