As a Network Engineer monitoring devices with heterogeneous sensor profiles,
I want to override the inherited Unusual Detection setting on an individual sensor (analogous to how channel limits can be set at sensor level),
so that I can silence chronic false-positive "Unusual" states on noisy low-value sensors (e.g., near-zero-traffic uplinks, msec-level Service sensors) without losing Unusual Detection on the device's other sensors and without splitting one physical device across multiple PRTG device objects.

Background:
Unusual Detection is currently inherited from Probe/Group/Device level, there's no per-sensor toggle. This becomes a problem when a device has a mix of sensors, some of which are good candidates for Unusual Detection and some that aren't, but you can't turn it off for just the noisy ones without also losing it on the useful ones.

Example: our core switch has 6 sensors — 5 are legitimate WAN/uplink traffic sensors with real day/night usage patterns where Unusual Detection is valuable (a sudden drop or spike is a real signal). The 6th is a near-zero-traffic uplink (<0.01 Mbit/s) where any normal jitter reads as a huge percentage swing, so it shows "Unusual" constantly even though nothing is wrong. The same issue shows up on low-execution-time Service sensors (msec-level execution times naturally jitter and trip the same false positive).

The current workaround is to move the noisy sensor onto its own dedicated device object and place that device in a separate group with Unusual Detection disabled, which means splitting one physical device's sensors across multiple PRTG device objects just to silence one sensor's false positives.

Request: add a per-sensor override for Unusual Detection (similar to how channel limits can already be set per-sensor), so a single chronically noisy sensor can be silenced without restructuring the device hierarchy or losing detection on the rest of the device's sensors.