why do some timer fleet commands run after reboot?

[rbucker]

I use fleet timers to replace crond. While I'm not certain that's the best tool for the job it's built-in and does it's own healing when there are runtime failures and so on. One odd occurrence is that some, and not all timer jobs, run after a reboot.

I'm looking at the systemd docs but have yet to determine the root cause. Whether it's specific to CoreOS or Systemd. Either way it should be documented.