Web applications are widely used by organizations to deliver some of their critical services to end-users. A key requirement is to ensure that the response time experienced by the end-user of Web service is always below a predefined threshold while having maximum throughput. This research work aims to ensure performance via efficient auto-scaling by utilizing tools from feedback control theory. Since there are no first-principles-based dynamic models for cloud computing systems, we emphasize on grey-box models. We use feedback control to calculate the appropriate number of Web servers required to serve the incoming Web requests while satisfying the response time criteria. Here, the feedback control computes and updates the control law at each sample instant. To avoid this, we utilize an event-triggered control framework where the control law gets updated as a consequence of an event. Also, we explore a data-driven control approach where the controller design does not require the explicit knowledge of system models. We perform real-time experiments to validate the efficacy and practicality of the proposed approaches.