Oil and gas refineries rely on scheduled maintenance windows to execute turnaround operations using custom-built applications hosted on Azure. These workloads exhibit unpredictable, bursty consumption patterns that challenge static scaling strategies, leading to cold-start latency, SLA breaches, and increased operational costs. This study addresses the lack of dynamic, real-time scaling frameworks tailored to refinery turnarounds by developing a combined Ansible-based automation solution for Azure App Service, databases and messaging queues enabling end to end scaling of azure services tuned to balance workloads. The proposed framework integrates infrastructure-as-code with continuous monitoring to enable agentless, adaptive scaling across services. Using a mixed-methods evaluation, we demonstrate that the hybrid strategy reduced median cold-start latency from 12.4s to 3.1s and cut SLA breaches from 18.7% to 2.3% under burst durations of 1–30 minutes. These improvements translated to a 27% reduction in resource waste and a 22% drop in cost per transaction during peak periods. The framework also enhanced reliability by minimizing manual intervention through automated configuration management. These results highlight how dynamic scaling can significantly improve performance and efficiency in industrial cloud environments. By integrating Ansible with real-time analysis of Azure service latency and coordinating End- to-End scaling across applications, databases, and messaging queues, the proposed approach offers a practical and cost-effective solution. This framework not only meets demanding performance requirements during volatile workloads but also provides a flexible model that can be adapted to other industries facing similar operational challenges.