Scaling Kubernetes Architectures for High Availability in Cloud

The shift from monolithic to microservices architectures has amplified the adoption of containerized applications, driving the need for efficient container orchestration platforms like Kubernetes. This paper presents a high-availability (HA) architecture for Kubernetes deployed on Amazon Web Services (AWS) to ensure seamless and resilient application performance in cloud environments. Key components of this architecture include load balancing, autoscaling, and centralized configuration management. The deployment leverages Amazon RDS for database reliability, Prometheus and Grafana for monitoring, and Horizontal Pod Autoscaling (HPA) to dynamically adjust resource allocation. Testing indicates that the system maintains high performance under various load conditions, with minimal downtime and rapid recovery from node failures. The proposed framework offers an adaptable and scalable solution for managing high-demand applications, providing a foundation for future optimizations under extreme user loads. The results shows that the system maintains a high success rate across various user loads and time intervals, with a minor drop under maximum load conditions. This behavior aligns with expected performance scaling in distributed systems, where resource allocation adjustments (Pods and worker nodes) enable high reliability across different user demands.