Hades Operator (Recommended)

The Hades Operator is the production-grade execution mode for Kubernetes. It implements the Kubernetes Operator pattern using Custom Resource Definitions (CRDs), giving Hades deep, native integration with the Kubernetes control plane.

Why the Operator?

Feature	Docker Executor	K8s Executor (deprecated)	Hades Operator
Kubernetes native	❌	Partial	✅
CRD-based job tracking	❌	❌	✅
Automatic retries	❌	❌	✅
Fine-grained RBAC	❌	Limited	✅
Reconciliation loop	❌	❌	✅
Helm chart support	✅	❌	✅

Architecture

The Operator introduces a BuildJob Custom Resource Definition (CRD). When the Scheduler receives a job, it creates a BuildJob resource. The Operator's controller loop watches for BuildJob resources and:

1. Creates a Kubernetes Job for each step.
2. Monitors Pod status and updates the BuildJob status accordingly.
3. Automatically retries failed steps based on the configured policy.
4. Cleans up completed resources.

NATS ──▶ Scheduler ──▶ BuildJob (CRD) ──▶ Operator Controller ──▶ Pod/Job
                                                │
                                                ▼
                                        Status & Logs

Configuration

The Operator mode is enabled by default when deploying with the Helm chart. The Scheduler must be configured with:

# helm/hades/values.yaml
hadesScheduler:
  executor: k8s
  configMode: serviceaccount

The serviceaccount config mode is the only supported mode. The Scheduler uses the pod's in-cluster ServiceAccount to authenticate with the Kubernetes API — no kubeconfig file is needed.

RBAC

The Helm chart automatically creates a ServiceAccount, Role, and RoleBinding that grant the Scheduler the minimum permissions required to manage BuildJob resources and monitor Pods within the hades namespace.

If you need cluster-wide access (e.g., to schedule jobs in multiple namespaces), set:

hadesOperator:
  clusterWide: true

This switches from a Role/RoleBinding to a ClusterRole/ClusterRoleBinding.

CRD Maintenance

The BuildJob CRD is defined in Go at:

HadesScheduler/HadesOperator/api/v1/buildjob_types.go

Whenever this file is modified, the generated files must be regenerated:

make -C HadesScheduler/HadesOperator manifests generate

This updates:

• HadesScheduler/HadesOperator/api/v1/zz_generated.deepcopy.go
• helm/hades/crds/build.hades.tum.de_buildjobs.yaml

A CI workflow verifies that committed generated files match the Go source on every pull request.

Submitting a Test Job

Apply the sample BuildJob manifest directly:

kubectl apply -f ./HadesScheduler/HadesOperator/config/samples/build_v1_buildjob.yaml

Monitor its progress:

kubectl -n hades get buildjobs
kubectl -n hades describe buildjob <name>