When you schedule a downtime over a specific tag scope, the downtime will apply "AND" logic to those tags. So if you wanted to set up a downtime over
host:B, then the downtime will silence only those monitors that alert on evaluation groups that contain both tags
host:B (and such a downtime would likely not silence any alerts).
If you want to schedule a downtime over a subset of hosts, a good way to do so is to find a host-tag common to all those hosts and scope the downtime by that tag. One easy way you can take advantage of to create new host-tags directly in the Datadog UI from your Infrastructure List and Host Map by selecting a host and hitting the "Edit Tags" button (shown below), or via our API.