Agility of DevSecOps processes: Dynamic log analysis approach

The ever-increasing need for flexibility and performance enhancement for the software industry has created the concept of DevOps today. Today, DevOps systems are of great importance for current software development processes. The significant increase in the importance of the security sector in recent years has also led to the emergence of the concept of DevSecOps. In this context, the autonomous process of the control, operation and development processes, which are constantly operating according to the unique structures of the systems, is of critical importance in order to provide flexibility and performance increase at low cost. Within the scope of this thesis, it is aimed to design autonomous DevSecOps processes in a more agile way. The proposed DevSecOps process aims to create system-specific security control mechanisms in order to develop software systems with the highest quality and security. The proposed system learns from the logs on the server for the autonomous security control mechanism. Within the scope of this learning, it gives statistical results in order to take precautions about the vulnerabilities that may be exploited on the system. The proposed DevSecOps model will work properly only if abnormal conditions and vulnerability characteristics are continuously detectable. In order to test the applicability of the DevSecOps model proposed in the study, log files that showing Docker resource usage statistics were created. By analyzing these log files, findings were created by analyzing the characteristics of abnormal situations and vulnerabilities with z-score and point biserial correlation methods. The result obtained according to the findings shows that it is possible to detect abnormal condition detection and vulnerability characteristics through logs.