RUM Intelligent Anomaly Detection¶
Leveraging intelligent detection algorithms, the system can automatically monitor performance anomalies and error spikes when users access websites or apps. By analyzing detection metrics, the system performs anomaly analysis and triggers alerts when necessary.
Performance analysis metrics include:
- LCP (Largest Contentful Paint ≥ 2.5 seconds);
- INP (Interaction to Next Paint ≥ 200 milliseconds);
- FID (First Input Delay ≥ 100 milliseconds);
- CLS (Cumulative Layout Shift > 0.01).
Based on these performance analysis metrics, the system calculates the percentage of affected users during the detection interval and triggers alerts according to predefined thresholds:
- Critical level: Affected user percentage exceeds 76.2%;
- Warning level: Affected user percentage exceeds 47.4%;
- Error level: Affected user percentage exceeds 59.8%.
Error detection metrics include:
- Session errors;
- View errors;
- Action errors.
Use Cases¶
Supports monitoring metric data anomalies for application types including Web and APP.
Configuration¶
- Define the monitor name;
- Select detection frequency: The execution frequency of the detection rule is fixed at 60 minutes;
- Select detection scope: Filter the data range to be detected through field filtering. If no filters are added, all application data will be detected.
View Events¶
The monitor retrieves metric information from the application service objects over the last 60 minutes. When anomalies are identified, corresponding events are generated. These anomaly events can be viewed in the Events > Intelligent Monitoring list.
Event Details Page¶
Click on Events to view the details page of an intelligent monitoring event, including event status, anomaly occurrence time, anomaly name, analysis report, alert notifications, history, and related events.
- Click Jump to Monitor in the upper right corner to view and adjust the Intelligent Monitor;
- Click the Export button in the upper right corner to choose between Export JSON File and Export PDF File, allowing you to obtain all key data corresponding to the current event.
Analysis Report
- Anomaly Summary: Displays RUM service labels, analysis report details, and anomaly distribution statistics;
- View Performance/Error Analysis: Links to the RUM View dashboard to analyze trends in metrics such as LCP, FCP, and view error rate.
Note
When multiple interval anomalies exist, the Anomaly Analysis dashboard defaults to displaying the first anomaly interval. You can click the "Anomaly Value Distribution Chart" to switch intervals; the anomaly analysis dashboard will sync accordingly.