In an environment with automatic updates enabled false alerts are created during reboots.

While it's possible to work around this by creating an elaborate custom script to enter maintenance mode it requires api credentials for zabbix on each of the hosts or some even more complex authentication method via another custom midware authentication module.

Another workaround would be creating a custom item with zabbix sender to notify of update process starting, but this would mean creating trigger dependencies which can also become a configuration hell over time.

A possible built-in solution for this could be to create a feature where the agent can be configured to receive a notice of an impending maintenance action and how long it's expected to last. Then the agent would forward this information to the server either directly or via proxy monitoring it. The server would be configurable using internal actions to add a host to a maintenance based on such a message arriving depending on what host sent it and how long the requested maintenance was.

So the flow would be:
1. Autoupdate command calls out zabbix agent runtime with maintenance command and duration
2. Zabbix agent sends maintenance request for it's related hostname to server/proxy
3. Server receives request for maintenance for a host
4. If such a host is configured under "automaintenance" rule under internal actions then relevant configuration is executed for the requested period.

Would such a feature be considerable? It seems like it could prove useful for a lot of users and give Zabbix a good new automation functionality.