Zabbix 3.0.10 after completion of the maintenance period 194 hosts with 97000 metrics, process timer led hosts from the "maintenance" and the process alerter began to send notifications OK on all triggers. I checked the table alerts in the beginning events with the symptom of the PROBLEM was not, generate only events are OK. The alerter process ran for more than 20 hours until it sent all notifications OK. While the queue is not gone, the notification of the last failures did not come. It is not clear why there was a mass generation of OK events on each trigger, after the maintenance period has ended?