SummaryCloud computing utilizes heterogeneous resources that are located in various datacenters to provide an efficient performance on a pay‐per‐use basis. However, existing mechanisms, frameworks, and techniques for management of resources are inadequate to manage these applications, environments, and the behavior of resources. There is a requirement of a Quality of Service (QoS) based autonomic resource management technique to execute workloads and deliver cost‐efficient and reliable cloud services automatically. In this paper, we present an intelligent and autonomic resource management technique named RADAR. RADAR focuses on two properties of self‐management: firstly, self‐healing that handles unexpected failures and, secondly, self‐configuration of resources and applications. The performance of RADAR is evaluated in the cloud simulation environment and the experimental results show that RADAR delivers better outcomes in terms of execution cost, resource contention, execution time, and SLA violation while it delivers reliable services.