The rapid adoption of the cloud-native apps has increased the demand of the distributed databases that may provide scalable, dependable and responsive data services to meet the dynamic computing environments. However, the unpredictable workload, resource contention, query performance, and data distribution, as well as the varying conditions of a cluster, also present an issue in deployments using Kubernetes. In this paper, a smart performance optimization system of distributed databases in cloud-based Kubernetes will be proposed based on AI. The framework is successful in embedding database monitoring, Kubernetes monitoring, workload-monitoring, AI-based optimization tools, and adaptive optimization systems within a single framework. It is also constantly gathering metrics concerning query latency, throughput, CPU and memory use, storage activity, network traffic and node performance in the database. These observations are used by the AI-based models to identify abnormalities in performance as well as their ability to predict the quantity of resources required, and suitable optimization actions. The framework helps to allocate resources dynamically, schedule allocations based on workloads, optimize query execution, manage replicas, and find intelligent locations of the data. The close loop control system enables continuous monitoring, decision making, optimization and validation with no manual intervention. The proposed architecture provides an organized remedy to the augmentation in the responsiveness, utilization of resources, scalability and economies of databases amid the varying workloads. The framework assessment reveals that it is a prospective smart base of autonomous and robust distributed database administration in the cloud-native infrastructures that employ Kubernetes
