一种结合知识图谱的告警收敛方法、系统
By introducing knowledge graphs and text similarity and time interval calculations, similar alarms in the database monitoring system are merged, solving the problem of alarm information overload and improving the accuracy of alarm processing and operational efficiency.
CN120104807BActive Publication Date: 2026-07-17BEIJING XINSHU TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING XINSHU TECH CO LTD
- Filing Date
- 2025-02-20
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies are insufficient to effectively address the high false alarm rate caused by the proliferation of alarm information in database performance monitoring systems, making it difficult for maintenance personnel to pinpoint the problem.
Method used
Knowledge graphs are introduced to represent the relationships between alarms. Alarm similarity is calculated by combining text similarity, time interval, and knowledge graph relationships. Similar alarms are merged through an alarm convergence method.
Benefits of technology
It improves the accuracy and efficiency of alarm convergence, reduces the number of alarms that maintenance personnel need to handle, and improves work efficiency.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN120104807B_ABST
Abstract
本发明提出一种结合知识图谱的告警收敛方法及系统，引入知识图谱表达告警之间的关联关系，并提出了一个结合文本相似度、时间间隔和知识图谱关联关系的告警相似度计算方法。通过这种方法，可以更加准确地判断告警之间的相似性，从而提高告警收敛的准确性和效率。本发明综合考虑告警内容、时间和知识图谱中的关联关系，对相似告警进行分组，有助于减少运维人员需要处理的告警数量，提高工作效率。由于使用了知识图谱中的复杂关联关系，因此易于深入理解告警之间的内在联系，这使得告警合并和收敛的过程更加精确，提高了告警处理的准确性。
Need to check novelty before this filing date? Find Prior Art