Microservices are now the most common architecture for designing software applications. Technology companies are also the target of an increasing number of cyberattacks on a daily basis, and new security solutions are in great demand. The use of intrusion detection system, that is characterized as the identification of irregular or unusual behaviour that occurs to a higher or smaller extent than typical instances in a data set, is one accessible measure. In this article, we extend previous work in which several real- world cyberattacks are launched against an operating microservices system, and the application activity is monitored and provided as distributed traces. A Stochastic Generative Adversarial Deep Convolution Infrastructure (SGADCI) is employed to model the collection of dispersed traces and to discover the geographical and temporal relationships of the data packets. Afterwards, the algorithm is used to produce predictions for current microservice activity, and cutoff point anomaly detection is employed to detect unusual microservice activity that indicates the presence of cyber security attacks that have been seeded, or anomalies. Cyberattacks such as brute force, batch activation of bot accounts, and disseminated denial of service were utilised to test this strategy.

cyber security, attack detection, microservices, cyberattacks, deep neural networks.