I have a puppet-ized Icinga installation across ~800 VMs. A handful of these are Windows. We use Puppet exported resources to provide host configs and a few other things. So our icinga2 server service restarts hourly with the new config. Whenever the Icinga2 server restarts, the windows agents lose connection for a few minutes and send out a notification. We have host checks every 10 minutes, and this issue never occurs with the Linux hosts. Only the Windows hosts.
Anyone else ran into hard state changes for Windows hosts only when the server restarts?