There was no real down-time, the service was actually online. The status monitor configuration was wrong which led to a in accurate "downtime" state.