Monitron Warning without breaking threshold

Amazon Monitron uses a combination of machine learning models and ISO vibration standards to identify warning and alarm conditions. Warnings can be triggered even when measurements don't appear to break visible thresholds because:

1. The system uses machine learning to detect early signs of potential failures based on patterns in the data, not just simple threshold crossings. During the initial learning period (14-21 days), Amazon Monitron establishes a baseline of "normal" operation for your specific asset.

2. After the learning period, Amazon Monitron continues to fine-tune this baseline and can detect subtle deviations from normal patterns that might not be obvious when looking at the raw data.

3. Warnings indicate that Amazon Monitron has detected early signs of a potential failure, while alarms indicate that measurements are consistently above thresholds and failure is more likely.

4. The system may be analyzing multiple factors simultaneously (temperature, vibration across different axes, patterns over time) rather than just the single measurement you're viewing.

Regarding total vibration versus single-axis vibration, the system does analyze these differently, and it's possible for patterns in the combined data to trigger warnings even when individual measurements appear normal.

Rather than moving thresholds up, it's recommended to acknowledge the warning, monitor the equipment closely, and investigate during your next planned maintenance. If you determine the equipment is healthy despite the warning, you can resolve the issue in the app with the appropriate action taken code, which helps the system continue to learn and improve its detection capabilities.
Sources
Understanding sensor measurements - Amazon Monitron
Understanding warnings and alerts - Amazon Monitron
Understanding asset status - Amazon Monitron
Amazon Monitron alerts - Amazon Monitron