Citation :

Radhakrishnan Pachyappan, Sarita Gahlot, Feroskhan Hasenkhan, 2025. "AI-Powered CI/CD Pipeline Optimization Using Reinforcement Learning in Kubernetes-Based Deployments" ESP International Journal of Advancements in Computational Technology (ESP-IJACT)  Volume 3, Issue 1: 132-139.

Abstract :

Continuous Integration and Continuous Deployment (CI/CD) pipelines are a fundamental part of modern DevOps, helping teams deliver software quickly and reliably. However, making these pipelines as efficient as possible is not an easy task. Challenges like poor resource allocation, deployment failures, and performance slowdowns in everchanging environments can hold teams back. This paper dives into how Reinforcement Learning (RL) can step in to tackle these challenges, especially in Kubernetes-based setups. RL agents learn from both past data and real-time information, making smart decisions to improve resource usage, speed up deployments, and handle failures more effectively. The paper takes a deep look at how RL works, compares it to traditional methods, and explores real-world examples. It also points to future research opportunities, showing how RL could revolutionize CI/CD pipelines by making them smarter, more adaptive, and capable of optimizing themselves over time.

References :

[1] Banala, S. (2024). DevOps Essentials: Key Practices for Continuous Integration and Continuous Delivery. International Numeric Journal of Machine Learning and Robots, 8(8), 1-14.
[2] Baitha, S., Soorya, V., Kothari, O., Rajagopal, S. M., & Panda, N. (2024, October). Streamlining Software Development: A Comprehensive Study on CI/CD Automation. In 2024 4th International Conference on Sustainable Expert Systems (ICSES) (pp. 1299-1305). IEEE.
[3] KAMBALA, G. (2023). Cloud-Native Architectures: A Comparative Analysis of Kubernetes and Serverless Computing.
[4] Tran, M. N., Vu, D. D., & Kim, Y. (2022, July). A survey of autoscaling in kubernetes. In 2022 Thirteenth International Conference on Ubiquitous and Future Networks (ICUFN) (pp. 263-265). IEEE.
[5] MUSTYALA, A. (2021). Dynamic Resource Allocation in Kubernetes: Optimizing Cost and Performance. EPH-International Journal of Science And Engineering, 7(3), 59-71.
[6] Flora, J., Gonçalves, P., Teixeira, M., & Antunes, N. (2022). A study on the aging and fault tolerance of microservices in kubernetes. IEEE Access, 10, 132786-132799.
[7] Santhanam, G. R. (2016). Qualitative optimization in software engineering: A short survey. Journal of Systems and Software, 111, 149-156.
[8] Sutton, R. S., & Barto, A. G. (1998). Reinforcement learning: An introduction (Vol. 1, No. 1, pp. 9-11). Cambridge: MIT press.
[9] Luz, H., Peace, P., Luz, A., & Joseph, S. (2024). Impact of Emerging AI Techniques on CI/CD Deployment Pipelines.
[10] Rayaprolu, R., Randhi, K., & Bandarapu, S. R. (2024). Intelligent Resource Management in Cloud Computing: AI Techniques for Optimizing DevOps Operations. Journal of Artificial Intelligence General science (JAIGS) ISSN: 3006-4023, 6(1), 397-408.
[11] AlMahamid, F., & Grolinger, K. (2021, September). Reinforcement learning algorithms: An overview and classification. In 2021 IEEE canadian conference on electrical and computer engineering (CCECE) (pp. 1-7). IEEE.
[12] Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347.

Keywords :

CI, CD, Kubernetes, RL, Automation, DevOps.