Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68878
Volume:
10 |
11 |
Nov 2026
Pages:
1-6
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.

Department of Computer Science and Engineering
Don Bosco Institute Of Technology

Department of Computer Science and Engineering
Don Bosco Institute Of Technology

Department of Computer Science and Engineering
Don Bosco Institute Of Technology

Department of Computer Science and Engineering
Don Bosco Institute Of Technology

Department of Computer Science and Engineering
Don Bosco Institute Of Technology
Workloads faced by cloud-native applications are extremely dynamic and unpredictable in nature; hence effective management of resources is a very important concern in today's era. Containers like Kubernetes have built-in capabilities of auto-scaling which help in managing workload of the applications. But the conventional methods of scaling mainly involve CPU thresholds which often might not react effectively towards unexpected traffic surges or any workload changes. The Proposed Custom V2 mechanism considers performance indicators: Requests Per Second (RPS), p95 response latency and CPU utilization to make the scaling decisions. The microservice model has been implemented using a Node.js based authentication microservice with REST API operations like user registration, login, forgot password, reset password and profile retrieval and integrated with MongoDB. The implementation also includes multiple security mechanisms, namely JWT authentication, bcrypt password hashing, Joi request validation and authentication middleware. The system has been benchmarked against the standard CPU-based Horizontal Pod Autoscaler (HPA) in Kubernetes to evaluate its performance, scalability and resource consumption for different workloads.
KEYWORDS
Cloud-native applications need good resource management to handle changing workloads. workload-aware autoscaling approach for a Node.js authentication microservice based on: requests per second (RPS), p95 response latency, and CPU utilization is proposed. The system combines Kubernetes, Prometheus and Grafana to monitor app performance and manage scaling decisions. The experimental results on Google Kubernetes Engine (GKE) show that the custom autoscaler achieved 0.9098 requests per second, while the CPU based HPA achieved 0.8916 requests per second, which is an improvement of 2.04%. The average latency reduced from 5.5625 seconds to 5.4394 seconds and the p95 latency reduced from 7.0425 seconds to 5.9913 seconds. The custom autoscaler also scaled replicas between one and five depending on the changes in the workload. Overall, these results indicate that the proposed workload-aware autoscaler can provide improved throughput and response-time performance compared with CPU-based HPA while dynamically adjusting application replicas according to workload variations.
[1]. Sultan Alharthi, “Auto-Scaling Techniques in Cloud Computing: Issues and Research Directions,” Sensors, vol. 24, no. 17, pp. 1–20, 2024, DOI: 10.3390/s24175551. [2]. Zhiqiang Zhou, Chaoli Zhang, Lingna Ma, Jing Gu, Huajie Qian, Qingsong Wen, Liang Sun, Peng Li, and Zhimin Tang, “AHPA: Adaptive Horizontal Pod Autoscaling Systems on Alibaba Cloud Container Service for Kubernetes,” arXiv, vol. 2303.03640, pp. 1–12, 2023, DOI: 10.48550/arXiv.2303.03640. [3]. Saurabh Agarwal, Rahul Kumar, and Ankit Singh, “A Deep Recurrent-Reinforcement Learning Method for Function Autoscaling,” arXiv, vol. 2308.05937, pp. 1–10, 2023, DOI: 10.48550/arXiv.2308.05937. [4]. H. Yuan, Y. Chen, and X. Zhang, “A Time-Series Forecasting Based Approach to Elastic Kubernetes Scaling,” Electronics, vol. 13, no. 2, pp. 1–14, 2024, DOI: 10.3390/electronics13020285. [5]. S. K. Mondal, S. K. Bera, and A. Ghosh, “On the Optimization of Kubernetes Toward Improved Cloud Resource Management,” Mathematics, vol. 12, no. 16, pp. 1–18, 2024, DOI: 10.3390/math12162476. [6]. V. Rajasekar, P. Kumar, and S. Baskar, “Security-Enhanced QoS-Aware Autoscaling of Kubernetes-Based Applications,” Journal of Intelligent Management Decision, vol. 3, no. 3, pp. 45–56, 2024. [7]. I. Pintye, B. Kovacs, and G. Szeidl, “Enhancing Machine Learning-Based Autoscaling for Cloud Applications,” Journal of Grid Computing, vol. 22, no. 4, pp. 1–16, 2024, DOI: 10.1007/s10723-024-09783-1. [8]. B. Jeong, “Autoscaling Techniques in Cloud-Native Computing,” Journal of Network and Computer Applications, vol. 220, no. 1, pp. 1–15, 2025, DOI: 10.1016/j.jnca.2025.103856. [9]. H. Guruge, P. B. Guruge, and S. Samarasinghe, “Time Series Forecasting Based Kubernetes Autoscaling Using Hybrid LSTM Models,” Frontiers in Computer Science, vol. 7, pp. 1–13, 2025, DOI: 10.3389/fcomp.2025.1509165. [10]. J. Santos, M. Ribeiro, and A. Costa, “Gwydion: Efficient Auto-Scaling for Complex Containerized Microservices,” Journal of Systems Architecture, vol. 150, no. 1, pp. 1–12, 2025, DOI: 10.1016/j.sysarc.2024.103210.
Publication Date
11/10/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
1
Total views since publication
PDF Downloads
5
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...