How a Silent 'n/a' Result Masked a Real Hardware Monitoring Bug for Months

A software developer discovered that a hardware health-monitoring script was returning a green 'OK' status on a Ryzen-based machine despite failing to read CPU temperatures, which were actually reaching 76°C. The bug stemmed from the script relying solely on Linux's thermal_zone interface, which Ryzen processors do not use, leaving the temperature variable at zero and triggering a silent 'not available' fallback. A built-in self-test routine technically caught the zero-sensor condition but labeled it a 'graceful off-node path' rather than raising a warning, effectively masking the failure as acceptable behavior. A separate monitoring tool running simultaneously confirmed the sensor data was accessible all along, reading the same die temperature to the degree via the hwmon interface. The author argues that an unresolved 'n/a' should never be treated as a passing result, and that monitoring tools must distinguish between a genuinely absent sensor and a failure to read one that exists.