If an alarm extension has setpoints outside the warning limits, the alarm extension shows as unreliable. The status reverts to the correct value when the setpoints are put within the warning limits. If this is happening on your system, you see AI points that were migrated from the N1 Network showing Unreliable on the system summary view at the NIE, even though they show as reliable at the OWS. The NIE focus view for the point then shows Reliable for the Reliability Flag and Unreliable for the Alarm State Flag.
The alarm extension generates the status; that is, if it is unreliable, then the status of the object it is attached to is also unreliable (even if the AI object itself is reliable). The alarm extension can go unreliable if the setpoint value is not between the warning limits.