DC Health: detecção de anomalias online em datacenters

Abstract

Datacenters are critical environments for the availability of technology-based services. Aiming at the high availability of these services, performance metrics of nodes such as Virtual Machines (VM) or VM clusters are widely monitored. Metrics such as CPU and memory utilization can show anomalous patterns associated with failures and performance degradation, culminating in resource exhaustion and total node failure. Early detection of anomalies patterns in data can enable remediation measures, such as VM migration and resource reallocation, before losses occur. However, traditional monitoring tools often use fixed thresholds for detecting problems on nodes and lack automatic ways to detect anomalies at runtime. In this sense, machine learning techniques have been reported to detect anomalies in computational systems with online and offline approaches. This work proposes the application called DC Health, as an approach to anticipate the online detection of anomalies in datacenter nodes. The purpose of DC Health is to detect anomalies in the behavior of hosts and alert datacenter operators so that investigation and remediation measures can be taken. This research was conducted from i) Systematic Literature Mapping, ii) problem modeling from real VM data and iii) DC Health evaluation using the prequential method in 6 datasets from the real world. The results showed that DC Health excelled in constant memory usage when processing data and detection accuracy in between 75% and 90%. As future work, it is mainly expected to evaluate the detection tool in cloud computing scenarios and develop automated mechanisms for diagnosis and remediation.