Writing cloud native applications today involves custom code to stitch everything together across microservices, open-source frameworks, and managed services. In this talk, we look beyond Kubernetes’ usage for container orchestration and extend the Kubernetes API to the management of external resources (not running on pods or nodes). Imagine if an application developer writing a monitoring framework could ask for a SQL database and message queue for their application in the same way that the Kubernetes API enables them to ask for an abstract PersistentVolumeClaim backed by a concrete PersistentVolume implementation. We can imagine that this developer might also expose the monitoring framework as a resource for consumption by other applications. In this talk we explore the use of this new approach to facilitate composability and reuse of resources within heterogeneous applications.
