I have a large number of virtual machines and laptops that aren't always kept on, making patch management of them difficult. Even installing the agent software doesn't help here, as I never know when the machine will become visible again to patch manager. This also negates scheduling, because although I could run a task (say) every hour looking for an anticipated machine, that wastes resources and there's no guarantee the machine will connect in the given time frame.
What I'd really like to do is to schedule tasks (especially patching + reboot) for "the next time this machine appears" even if this requires the installation of the software agent. I could then rest easier knowing that the next time the software agent connected to the patch manager, it would receive and perform task instructions queued in a schedule. The schedule model still works here as it makes it easy to disable, modify, or delete the queued actions. The tricky bit would be resolving the notification actions after a machine in the target group was updated. For that, I'd like to see a new notification check boxes along the lines of "Send email after each machine action" and "Send email after schedule completes". That would allow me to know when each machine was handled and when the schedule was finally complete. The current schedule time limitations might also be used to limit the schedule to being "active" for a limited time, so that the queued actions last only (say) a month, before they're obviated.
The feature would truly be a godsend, because the manual overhead to continually hunt down machines is noticeable, as is the pie chart showing that we're never really 100% complete. Of course I can (and do) delete machines that don't show for a while, but as soon they re-appear the metrics skew. Unfortunately they can also appear and disappear within a few hours, so we don't always get the chance to manually push actions to them. This feature would fix all of that.
Thanks