Mitigasi Bottleneck Tingkat Koneksi pada Kubernetes HPA menggunakan Metrik RPS untuk Layanan Nginx Reverse Proxy
DOI:
https://doi.org/10.55606/juisik.v6i2.2397Keywords:
Autoscaling, Bottleneck, Horizontal Pod Autoscaler, Kubernetes, Requests Per SecondAbstract
Default CPU metrics in the Kubernetes Horizontal Pod Autoscaler (HPA) often fail to detect connection bottlenecks in I/O-bound workloads, such as Nginx reverse proxies, causing service failures despite low CPU utilization. This study aims to evaluate and compare the effectiveness of CPU-based HPA against custom Requests Per Second (RPS) metrics in mitigating connection-level bottlenecks. An experimental approach was conducted using a two-tier architecture (Nginx and a backend with a 500ms delay) under ramp-up and sustained high-load traffic scenarios. The results demonstrated that the CPU-based HPA stagnant with low CPU usage, producing HTTP 5xx error rates up to 57.4% during ramp-up and stagnating at 36% under sustained load due to its inability to detect operating system-level network queues. Conversely, the RPS-based HPA proactively mitigated these bottlenecks by executing timely scaling, suppressing error rates below 10% during ramp-up and significantly restoring service stability. These findings imply that utilizing RPS metrics successfully decouples autoscaling triggers from hardware resource saturation levels, making it a highly reliable metric for maintaining high availability in connection-limited microservice architecture.
References
Ahmad, H., Treude, C., Wagner, M., & Szabo, C. (2024). Smart HPA: A resource-efficient horizontal pod auto-scaler for microservice architectures. In 2024 IEEE 21st International Conference on Software Architecture (ICSA) (pp. 46–57). IEEE. https://doi.org/10.1109/ICSA59870.2024.00013
Camilo, J., & Campos, N. (2024). Evaluating the transition from HPA to KEDA in an event-driven architecture deployed in Kubernetes.
Gao, R., Xie, X., & Guo, Q. (2017). K-TAHP: A Kubernetes load balancing strategy based on TOPSIS+AHP. XX.
Gawande, S., & Gorthi, A. (2024). Containerization and Kubernetes: Scalable and efficient cloud-native applications. International Journal of Innovative Science and Research Technology (IJISRT), 435–439. https://doi.org/10.38124/ijisrt/IJISRT24NOV314
Guedes, E. A. C., Silva, L. E. T., & Maciel, P. R. M. (2024). Performability analysis of I/O bound application on container-based server virtualization cluster.
HPA Documentation. (n.d.). Horizontal Pod Autoscaler. Retrieved March 29, 2025, from https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/
Huotari, H. (2025). NGINX as an API gateway.
Jónsson, A. J. (2020). Hybrid auto-scaling for an asynchronous computationally intensive application [Degree project, Computer Science and Engineering].
Campbell, J. (n.d.). Kubernetes vs. Docker: A comprehensive guide to containerization. Atlassian. Retrieved March 29, 2025, from https://www.atlassian.com/microservices/microservices-architecture/kubernetes-vs-docker
Kubernetes Documentation. (n.d.). Kubernetes documentation. Retrieved February 26, 2025, from https://kubernetes.io/docs/home/
Lorido-Botran, T., Miguel-Alonso, J., & Lozano, J. A. (2014). A review of auto-scaling techniques for elastic applications in cloud environments. Journal of Grid Computing, 12(4), 559–592. https://doi.org/10.1007/s10723-014-9314-7
Malallah, H. S., Qashi, R., Abdulrahman, L. M., Omer, M. A., & Yazdeen, A. A. (2023). Performance analysis of enterprise cloud computing: A review. Journal of Applied Science and Technology Trends, 4(1), 1–12. https://doi.org/10.38094/jastt401139
Molleti, R. (2022). Kubernetes advanced auto scaling techniques. Journal of Mathematical & Computer Applications, 1(4), 1–4. https://doi.org/10.47363/JMCA/2022(1)E126
Nilsen, J. (2023). Performance evaluation of Kubernetes autoscaling strategies on GKE clusters.
Noor, J., Faysal, M. D. B., Amin, M. D. S., Tabassum, B., Khan, T. R., & Rahman, T. (2025). Kubernetes application performance benchmarking on heterogeneous CPU architecture: An experimental review. High-Confidence Computing, 5(1), Article 100276. https://doi.org/10.1016/j.hcc.2024.100276
Raj, P., Vanga, S., & Chaudhary, A. (2023). Kubernetes architecture, best practices, and patterns. In Cloud-native computing: How to design, develop, and secure microservices and event-driven applications (pp. 49–70). Wiley. https://doi.org/10.1002/9781119814795.ch3
Šimon, M., Huraj, L., & Búčik, N. (2023). A comparative analysis of high availability for Linux container infrastructures. Future Internet, 15(8), Article 253. https://doi.org/10.3390/fi15080253
Microsoft Azure. (n.d.). What is cloud computing? Retrieved February 26, 2025, from https://azure.microsoft.com/en-us/resources/cloud-computing-dictionary/what-is-cloud-computing/
Wirawan King, R. Z., Hari Trisnawan, P., & Yahya, W. (2024). Perbandingan metode autoscaling vertical pod autoscaler dan horizontal pod autoscaler Kubernetes pada Google Cloud Platform. Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer.
Zhou, H., & Yong, C. H. (2024). Implement HPA for Nginx service using custom metrics under Kubernetes framework. IEEE Access, 12, 189722–189734. https://doi.org/10.1109/ACCESS.2024.3509876
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Jurnal ilmiah Sistem Informasi dan Ilmu Komputer
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
