Root Cause Analysis Bot using Machine Learning Techniques Tapan Kumar Behera 1 and Kumud Tripathi 2 1 Technical Architect, Cambridge, MA, USA 2 Department of CSE, IIIT Lucknow, India Corresponding author: Tapan Kumar Behera, E-mail: tkbehera.usa@gmail.com November 11, 2022 Abstract In this world of quick delivery of quality products, DevOps test automation plays a vital role. This triggers the automation build to test whether the product quality is a good fit for the release or not? When multiple test cases of a test suite got failed, it takes lots of time for the developer and analyst to analyze the error logs of each test case and do the root cause analysis (RCA). For one test case if RCA takes around 20 minutes then for 100 test cases it will be going to take around 33 hours and it’s nothing but 4 Analysts’ full-day work, then it’s a question on ROI(Return on Investment) ? To solve this time-consuming process, we are introducing Root Cause Analysis Bot (RCA Bot) in this paper. This bot will analyze the failure logs, error message, descriptions, error codes and apply the machine learning (ML) techniques to predicts the root cause of the failure with a percentage of the prediction accuracy. Index terms— Machine learning, Artificial Intelligence, Text Processing, Natural Language Processing, Multi-class Classification, Spark, Scala, System Architecture, Software Applications, En- terprise Application, Distributed Application, Microservice, DevOps 1 Introduction Enterprises with complex distributed applications designed with multi-tier architecture systems and with a lot of third-party applications communicating each other, cannot prevent failures from hap- pening. As a result, finding the cause of failures becomes a real challenge, as it requires cross-team cooperation, log analysis, and many more., and becomes part of the daily routine, hindering the organization’s productivity. Second, since a product’s complexity and a short time to market make it difficult to maintain its quality, it is also important to keep the cost low to remain competitive in the market. Software devel- opers must consider many aspects when developing any product, which makes it essential to take this into consideration. Despite following all guidelines, software defects still occur. Many factors can lead to defects and failure of the product, such as a human factor, lack of communication, poor design logic, unrealistic development deadlines, untrained testers, and poor coding practices[21]. Furthermore, the 1