Abnormal event identification and alerting method, apparatus, device, and medium
By integrating log collection, cleaning, parsing, indexing, analysis, display, and alarm processing, the problem of lagging log data processing across multiple data nodes has been solved, achieving timely and accurate anomaly identification and intuitive risk monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- PING AN HEALTH INSURANCE CO LTD
- Filing Date
- 2026-04-14
- Publication Date
- 2026-06-26
AI Technical Summary
Existing technologies struggle to collect, structure, and analyze log data from multiple data nodes in a unified manner, resulting in delayed anomaly identification, unintuitive display, and untimely alarm response.
By connecting multiple data nodes through log collection tools, raw log information is obtained, fields are cleaned and structured parsed to generate structured log data, and a query index is built on the data processing platform. Anomalies and their severity are analyzed using a predictive model, visualization components are called to generate and display information, and alarm information is generated and sent to the receiving terminal.
It achieves timely and accurate anomaly identification, improves the intuitiveness and response efficiency of risk monitoring, and is applicable to business scenarios such as fintech and healthcare.
Smart Images

Figure CN122285443A_ABST