
软件学报 ISSN 1000-9825, CODEN RUXUEW
Journal of Software, [doi: 10.13328/j.cnki.jos.006177]
©中国科学院软件研究所版权所有.
金连源
,
李国良
(清华大学 计算机科学与技术学院,北京 100084)
通讯作者: 李国良, E-mail: liguoliang@tsinghua.edu.cn
摘 要:
始关注数据库运行的稳定性.
而这可能会带来巨大的经济损失.
标有数百个,
这种成本会是很多公司难以接受的.因此,
如何用较低的成本完成对数据库的自动监控和诊断是一个具有挑战性的
问题.现有的 OLTP
数据库自动异常诊断方法往往存在着监控信息收集成本过高
问题.在这篇论文里我们提出了一种智能
取和根因分析这三个模块,这三
验、和优化的 K 近邻算法.
据库,
关键词: OLTP 型数据库;异常诊断;人工智能
中图法分类号: TP311
AI-Based Database Performance Diagnosis
JIN Lian-Yuan, LI Guo-Liang
(Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China)
Abstract:
Database is a kind of important and fundamental computer system software. With the
all walks of life, a growing number of people
begin to concern the stability of the database. Because of the numerous internal of
effect, performance anomaly may emerge when the Database running and it
database anomaly by analyzing monitoring metrics. However, there are hundreds of metrics in the system and
unable to extract valuable
information from them. Some major companies employ DBA to manage the
unacceptable for many other companies. Achieving automatic
database monitor and diagnose with low cost is a challenging problem.
Current methods have many limitations, including high cost of metrics information collection, narrow
In this paper, we propose an anomaly diagnose framework
contains LSTM-based anomaly detection
module and modified
framework consists of an offline training and an online diagnose stage. Our
diagnose accuracy with minor overload to system performance
Key words: OLTP Database; Anomaly Diagnosis; Artificial
IT 运维,指的是和 IT
且使服务达到用户付款后的预期.
500 亿到 1000 亿的 IT 设备,
基金项目: 国家自然科学基金(61925205, 61632016)
Foundation item: National Natural Science Foundation of China (61925205, 61632016
收稿时间: 2020-07-19; 修改时间: 2020-09-03; 修改时间
E-mail: jos@iscas.ac.cn
http://www.jos.org.cn
Tel: +86-10-62562563
,随着数据库在各行各业的广泛应用,越来越多的人开
,数据库在实际运行的过程中会出现性能异常,
,但是关于数据库监控指
.一些传统的公司会聘用专业的人员管理数据库,而
如何用较低的成本完成对数据库的自动监控和诊断是一个具有挑战性的
数据库自动异常诊断方法往往存在着监控信息收集成本过高
、适用范围小抑或是稳定性较差等
AutoMonitor,提供了数据库异常监测、异常指标提
LSTM 的时间序列异常诊断模型、Kolmogorov-Smirnov 检
.我们将提出的系统部署在 PostgreSQL 数
,并且不会对系统性能造成太大的影响.
(Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China)
Database is a kind of important and fundamental computer system software. With the
development of database application in
begin to concern the stability of the database. Because of the numerous internal of
external
effect, performance anomaly may emerge when the Database running and it
may cause huge economic loss. People usually diagnose
database anomaly by analyzing monitoring metrics. However, there are hundreds of metrics in the system and
ordinary database users are
information from them. Some major companies employ DBA to manage the
databases but this cost is
database monitor and diagnose with low cost is a challenging problem.
Current methods have many limitations, including high cost of metrics information collection, narrow
range of application or poor stability.
which deployed on the PostgreSQL database. The framework
K Nearest-Neighbor algorithm-based root cause diagnose module. The
framework consists of an offline training and an online diagnose stage. Our
evaluations on the datasets show that our framework has high
,IT 运维可以让公司提供的服务保持良好的质量,并
.预计到 2020 年全球将有
,覆盖互联网、金融、物联网、智能制造、电信、电力
Foundation item: National Natural Science Foundation of China (61925205, 61632016
)
-11-06; jos 在线出版时间: 2021-01-20
评论