Werk #8677: Fixed persistent recurring downtimes during cofiguration reload

The Check_MK Micro Core did not handle persistent recurring downtimes correctly during configuration reloads: Already existing ones were removed and re-added, but there was a small time window where the downtime was not actually active (between the reload and the next time all downtimes are handled, i.e. up to 3 seconds). This could lead to notifications during downtimes, e.g. when a host or service was in a non-OK state during the reload.

To the list of all Werks