Forecasting and Anomaly Detection
AIOps for NGFW
can use machine learning
to expose unusual behavior in your deployment.Forecast and Anomaly Alerts are available
only with a Premium
AIOps for NGFW
license.Normally,
AIOps for NGFW
detects issues by applying fixed rules to the metrics
in your deployment. For example, if Management Plane
CPU Usage
exceeds 85%, the metric enters a Critical
state.However, to alert you to events that fixed rules might
miss,
AIOps for NGFW
can use machine learning to understand
your deployment and offer you additional alerts tailored to your
usage trends.- Forecast-Based Alertshelp you anticipate issues by projecting how a device metric may change and alerting you accordingly.
- Anomaly-Based Alertsestablish a baseline behavior for a device metric and alert you when that metric crosses anAnomaly Thresholdthat you specify.
- Enable from .
- Set theDefault Anomaly Threshold.This is the number of standard deviations, from 1 to 50, that determines how much a metric can change from its baseline before it triggers an alert.
- Select an alert that you want to investigate for anomalous behavior.You can also change theAnomaly Thresholdfrom within the alert. This will override the global setting for this alert.
- Use chart tools to inspect anomalous periods.
TheNormality Bandshows you the range determined by theAnomaly Thresholdthat you set.Threshold Eventsshow you precise moments when anomalies occurred.
