DevOps for IoT Systems: Fast and Continuous Monitoring Feedback of System Availability

Current Internet-of-Things (IoT) systems are highly distributed systems, which integrate cloud, fog, and edge computing approaches. Accelerating their maintenance and continuous improvement, while ensuring their availability, is complex. DevOps promotes fast and continuous feedback from operations to development to detect problems before customers are impacted, among other benefits. However, there is not any formal definition of how to do this. This article defines the “fast and continuous monitoring feedback of system availability” activity (F&CF availability) that supports automatic and continuous monitoring feedback from operations to the development of the IoT system availability. This activity has been formalized through the software and systems process engineering metamodel (SPEM). Its implementation is demonstrated in a real scenario that provides evidence that the formalization of the F&CF availability activity helps teams in better diagnosing and fixing outage problems. The result is a distributed and configurable monitoring component developed through code [monitoring as code (MaC)]. This component is embedded in the IoT infrastructure. MaC enables DevOps team to configure their own metrics and indicators at runtime, i.e., monitoring on demand. The formalization of this activity, based on an MaC technique, enables the automation, versioning, and replication of monitoring elements.