Ever have one of those bugs that just drives you crazy? I squashed one today that had been haunting me for awhile.
It involved a node losing a service, followed immediately by the node being marked down. The node would then be marked up and the original service that failed would remain down.
This was especially painful when it happened during one of my training classes. Since apache is almost always installed on Red Hat machines I come across, but not running, it is usually a nice exercise to have students start the httpd service, perform a rescan, and see the service added to the WebUI.