It seems the maintenance period is interrupted by daylight saving from CEST to CET. Hosts which where in maintenance mode without data collection, where monitored and triggered in a timeframe between 31.10.2021 23:00 and 31.10.2021 23:59. Because of this we got a lot of false alarms, which trigger our emergency response team. Since we had daylight saving on 31.10.2021 at 03:00 this day had 25 hours. I guess this is the reason our maintenance period was interrupted for 1 hour.
As you can see in the screenshots we have a special host group for hosts in maintenance mode for an unknown period. Thus this maintenance period should last "for ever" as long as this host is in that host group.

Steps to reproduce:

1. Create host with icmp ping item and trigger, which is activly monitored
2. Block icmp to that host or switch the host off, wait for trigger.
3. Create a maintenance window as in the screenshots on a daily base and a schedule of 0:00 every day, lasting for 1 day.
4. Put host into "maintenance" host group
5. Create a trigger action for this host if not reachable. Pause operations for suppressed problems.
6. Set system time to 30.10.2021 23:59 CEST
7. On 31.10.2021 03:00 CEST the System will switch to 31.10.2021 02:00 CET
8. Wait until 31.10.2021 23:00 CET

Result:

This host in maintenance host group will be monitored again. It is out of maintenance mode. The trigger action of this host will be fired.
See the screenshot of the ping history graph.

Expected:
Systems in maintenance host group shall not be monitored and trigger no alarms.