This chapter introduces intelligent autoscaling in Kubernetes using the MAPE (Monitor-Analyze-Plan-Execute) framework and its extended form, MAPE-K, which incorporates knowledge-driven decision-making. It begins with an explanation of MAPE components and their application in reactive autoscaling, highlighting both the advantages of responsiveness and the limitations of delayed reaction times. The proactive approach within the MAPE framework is then discussed, demonstrating how predictive analysis and planning can anticipate workload changes for optimized scaling. The MAPE-K extension is examined in terms of its uses, benefits, and challenges, particularly its role in adaptive, knowledge-based resource management. Performance evaluation is supported by key metrics, along with real-world implementations and case studies that show practical applications of the framework. The chapter concludes with a discussion of advanced research directions in intelligent autoscaling, emphasizing how MAPE and MAPE-K approaches can enhance cloud-native and edge computing environments by unifying automation, prediction, and adaptive orchestration.