Self-healing Mechanisms: Analysis and Implementation for Stringent-SLAs Products

This thesis explores self-healing mechanisms in cloud applications, with a focus on reducing manual intervention in system reliability engineering (SRE). Conducted as part of an internship at Amadeus, the study examines existing self-healing solutions, evaluates their integration within the company’s technology stack, and proposes improvements to automated incident remediation. The research covers built-in self-healing features of key infrastructure components such as Couchbase, Oracle DB, Kubernetes, and cloud platforms. Additionally, it investigates external monitoring and automation tools, including ARGOS, ServiceNow, and AWX, to enhance alert management and remediation processes. A key outcome of this work is the enhancement of Amadeus’ "Auto Remediation of Alerts" framework, improving its ability to autonomously resolve common incidents. The findings contribute to the broader field of SRE by showing how automation can improve system availability, reduce operational costs, and enhance response times to failures.