If an alarm (or warning or whatever) is raised, something bad has happened.
Testing that the condition was properly flagged is a key consideration
Anomalous Condition
Did the system catch the anomalous condition and properly notify the user?
But what happens after that is just as important.
Does the alarm persist if the condition is not resolved?
Does the alarm time out if ignored altogether?
Take Appropriate Actions
Regardless, when we drive the system to raise an anomalous condition, we always want to follow up the test to demonstrate that it properly recovers from the anomalous condition when the user takes the correct and appropriate actions.