Currently we have several instances grouped into three top-level groups and then several subgroups underneath. This structure allows us to set alerts at the group level and then drop instances in and out of that defined set.

If we select a group and configure an alert setting, when the page refreshes it does not return back to that same sub-group, instead opting to go back to the very top 'All Servers' level which if not spotted can mean you then select the next alert to configure and inadvertently change the setting for all servers.

This seems like basic functionality but I cannot find anywhere else to raise this as its clearly a design choice instead of a bug, just a bad one in my opinion.

In a related note, its not always clear which instances have 'overridden' rules and the bold type face doesn't always appear if the alert is on a job or disk buried several layers deep. It would be nice to see a way to show "only overridden alerts" or something as a filter.