Anomaly Alerts

Anomaly alerts are automated monitoring systems that detect deviations from established normal patterns in network operations. They employ statistical analysis to create baseline thresholds through a 14-day learning period, analyzing metric distributions (min/max values, standard deviation) across different time intervals. They then determine the parameters for an alert.

These alerts enable you to identify potential issues in network infrastructure before they escalate. Anomaly alerts are particularly useful for detecting zero-day anomalies and subtle performance degradation.

After the learning phase, violations trigger when metrics exceed statistically calculated ranges.

Key characteristics:

• Automatic Thresholding: Eliminates manual configuration by learning normal patterns
• Pattern Recognition: Analyzes daily/weekly cycles and seasonal variations
• Proactive Detection: Flags unusual activity like traffic spikes, resource anomalies, or performance deviations
• Multi-Metric Analysis: Processes 50+ device metrics simultaneously

Enabling a device

1. Select the device in the [Inventory] tab.

2. In the [Monitors] tab, double click to select the incident to which the anomaly alert will be applied.

3. Click the [Triggers] button.

4. Add in your message

5. Click [Save], then [Close].

This will cause the anomaly alert to run for 14 days, during which time it will learn the parameters to alert on.