Before applying windows updates to our servers we put them into SCOM maintenance mode using the SCOM console, this works well for most of our servers except SQL and Exchange 2010 servers where a large amount of alerts sent during reboot, some examples as follows:

• Cluster Node is not online. on Not Present
• Cluster resource group offline or partially online on Not Present
• The Windows Remote Management Service (WinRM) isn't running.
• Failed to Connect to Computer on Microsoft.SystemCenter.AgentWatchersGroup
• The database availability group is experiencing problems that may cause it to fail due to stale security information

I'm putting the windows computer, cluster nodes, and DAG into maintenance mode, but obviously something else has to be put into MM?

Hi, it's not difficult to identify from what objects you receive these alerts (in properties) to take it into account

and some inf.

http://www.opsman.co.za/scom-cluster-maintenance-mode-resource-groups-and-you/

http://syscenterstuff.blogspot.ru/2010/02/maintenance-mode-clusters-nodes-sql.html

http://blogs.msdn.com/b/mariussutara/archive/2008/09/05/cluster-and-maintenance-mode.aspx

Agreed. If you are still getting Alerts, take a look at the Alert property (specifically the Target), which will let you know which object it is being thrown from; and from there, which object you have not put into Maintenance Mode. 

Part of the problem with the DAG alerts is that they're generated by the Exchange Correlation Engine, so the rule target is the Root Management Server. DO NOT EVER EVER EVER put the RMS into maint mode.