On translating monitoring insights into a cloud-native execution environment

Software systems are getting more advanced, resulting in the need for automated management. It especially applies to Cloud-native applications consisting of numerous microservices. Typically, such a service generates large amounts of logs and system metrics. Manually analyzing them is tedious and results in a delayed reaction. To overcome the obstacles, as our unique contribution we propose translating the metrics into objects to allow for rule-based management and meanwhile benefit from object-oriented programming. By leveraging metrics data, we can define rules for the automatic handling of certain situations. Our methodology is based on an experimental evaluation of the proposed concept. We named our proposal Hephaestus. We also present the reasoning behind its design and show a real-life scenario in which Hephaestus can be helpful. In conclusion, Hephaestus introduces a new telemetry objects layer in a Cloud-native environment. In this layer, the telemetry data exist as objects. These objects are understandable by various software that, based on them, can enhance the insights.