Formulating an Korean LLM-Based Interactive Assistant for Enhanced IT Collaboration in Microservice Environments

This research explores container metrics in microservices, implement an LLM-based IA to leverage the principles of Large Language Models (LLMs) to gain insights into these metrics. This IA is designed to analyze metrics, provide actionable insights, and foster efficient communication between IT operators and microservice systems. Using a set of diverse and representative workload data, the study conducted realistic workload injection experiments to gather performance metrics for each constituent microservice. The collected data was then analyzed and written the insight of that into the natural language. The LLM-based IA model uses this information to comprehend microservice structures and workload characteristics. Consequently, the proposed IA model demonstrates superior question-answering performance in the microservices-specific domain compared to the ChatGPT model, which serves as a benchmark in this study.