Why Orkestra

2 min read

Building a Kubernetes operator means writing Go: controllers, informers, reconcilers, finalizers, events, metrics. Every team does this. Every team builds the same infrastructure. The only part that differs is the business logic — and it is usually just 10 lines of YAML describing what you want.

Orkestra inverts this. You write a Katalog — a declarative YAML file that describes your CRDs and what should happen when a Custom Resource is created, updated, or deleted. Orkestra provides the runtime: informers, worker pools, drift correction, finalizers, events, metrics, health endpoints.

No Go. No code generation. No controller boilerplate.

What You Write

apiVersion: orkestra.orkspace.io/v1
kind: Katalog
metadata:
  name: webapp-operator
spec:
  crds:
    webapp:
      crdFile: ./crd.yaml
      operatorBox:
        default: true
        onCreate:
          deployments:
            - image: "{{ .spec.image }}"
              replicas: "{{ .spec.replicas }}"
              reconcile: true
          services:
            - port: "80"
              targetPort: "{{ .spec.port }}"
              reconcile: true

That is a fully working operator. Orkestra reads the CRD from crdFile, applies it to the cluster, starts informers, and reconciles every WebApp CR into a Deployment and Service. Drift correction included.

What Orkestra Does

When you run ork run:

  1. Reads your Katalog and finds crdFile: ./crd.yaml
  2. Applies the CRD to the cluster automatically
  3. Starts informers watching for CRs of that type
  4. For each CR: resolves templates, creates resources, adds finalizers
  5. On every reconcile: corrects drift on resources marked reconcile: true
  6. On CR delete: cleans up all child resources

All of this from a YAML file.

Who It Is For

  • Platform engineers building internal developer platforms without writing operators
  • SREs who want declarative infrastructure, not Go code
  • Application teams defining their own CRDs without learning controller patterns

Get Started

curl -sSL https://get.orkestra.sh | bash
ork init
ork run
# Orkestra reads katalog.yaml from the current directory and starts the runtime.

→ Getting Started

Further Reading