Alex.Eze

Separate reconciliation behaviour from operator infrastructure.

Kubernetes already provides a battle-tested reconciliation model, but using it traditionally means building and maintaining controllers, watches, queues, retries, status handling, and cleanup around application-specific logic. That machinery is often repeated even when the decision being made is completely different.

Inrun makes the machinery reusable. Supply an intent and implement the decision using declarative rules, a remote HTTP endpoint, or typed Go reconciliation. Inrun prepares the request, executes the behaviour, applies permitted resources, and manages the reconciliation lifecycle. The caller-facing intent can remain separate from the internal CRD representation.

Kubernetes Expression Library

KEL is the reusable expression-function layer behind Inrun's templates: pure, nil-safe helpers for inspecting Kubernetes objects, working with fields and collections, validating values, and expressing common operational logic. The aim is to make that Kubernetes knowledge useful outside the Inrun runtime too.