ZTE Communications ›› 2022, Vol. 20 ›› Issue (3): 85-92.DOI: 10.12142/ZTECOM.202203011

• Research Paper • Previous Articles     Next Articles

Approach to Anomaly Detection in Microservice System with Multi- Source Data Streams

ZHANG Qixun1, HAN Jing2(), CHENG Li2, ZHANG Baisheng2, GONG Zican2   

  1. 1.Peking University, Beijing 100091, China
    2.ZTE Corporation, Shenzhen 518057, China
  • Received:2022-01-24 Online:2022-09-13 Published:2022-09-14
  • About author:ZHANG Qixun is currently an assistant professor in School of Software and Microelectronics in Peking University, China. He received his PhD in 2022. His research interests include distributed systems, AIOps, etc.|HAN Jing (han.jing28@zte.com.cn) joined ZTE Corporation in 2000. She is an expert in AIOps. She has been putting effort into natural language process for over 10 years and has published several papers.|CHENG Li joined ZTE Corporation in 2006. He is an expert in AIOps and wireless communications. He has much experience in analyzing variety of types of data. He has a lot of experience in problem-solving and methodology.|ZHANG Baisheng joined ZTE Corporation in 2011. His work has been devoted to cell-phone terminal techniques for over 10 years. Besides, he is interested in the research of auto-driving technology.|GONG Zican joined ZTE Corporation in 2020. He received his master’s degree in computing from Australian National University, Australia in 2019. His research interests include AIOps and natural language processing.
  • Supported by:
    ZTE Industry-University-Institute Cooperation Funds(HF-CN-202008200001)

Abstract:

Microservices have become popular in enterprises because of their excellent scalability and timely update capabilities. However, while fine-grained modularity and service-orientation decrease the complexity of system development, the complexity of system operation and maintenance has been greatly increased, on the contrary. Multiple types of system failures occur frequently, and it is hard to detect and diagnose failures in time. Furthermore, microservices are updated frequently. Existing anomaly detection models depend on offline training and cannot adapt to the frequent updates of microservices. This paper proposes an anomaly detection approach for microservice systems with multi-source data streams. This approach realizes online model construction and online anomaly detection, and is capable of self-updating and self-adapting. Experimental results show that this approach can correctly identify 78.85% of faults of different types.

Key words: anomaly detection, data stream, microservice, monitored indicator, system log