Sr Software Developer at MongoDB
Seasoned software developer with plenty of years experience in the field. Passionate about technology and distributed systems in particular, such as Kubernetes. I have contributed to industry leaders including AWS, Canonical, Bitnami, and VMware. Currently, I focus on building scalable data infrastructure at MongoDB.
Traditional Kubernetes Operator development often starts with code, leaving the user interface—the CRD—as an afterthought. This talk proposes a declarative, schema-first workflow that flips the script: defining the CRD interface first and deriving the Go types and controller logic from it.
By treating the OpenAPI specification as the source of truth, we can automate the boilerplate away. This approach standardizes (or "regularizes") the codebase, moving away from the sprawling, custom logic often found in a standard Reconcile() loop. Instead, we model resource behavior under a universal state machine.
Using the MongoDB Atlas Operator as a case study, we will demonstrate how this transition from "custom-coded" to "schema-driven" results in a more maintainable, predictable, and scalable operator.
