This repository follows GitOps principles for managing OpenShift AI infrastructure and configurations.
openshift-ai-gitops/
├── README.md
├── base/
│ └── README.md
├── components/
│ └── README.md
├── services/
│ └── README.md
└── dependencies/
├── kustomization.yaml
├── cluster-config/
│ ├── kustomization.yaml
│ └── default-kueue.yaml
└── operators/
├── kustomization.yaml
├── README-labels.md
├── cert-manager/
│ ├── kustomization.yaml
│ ├── namespace.yaml
│ ├── subscription.yaml
│ └── operatorgroup.yaml
├── cluster-observability-operator/
│ ├── kustomization.yaml
│ ├── namespace.yaml
│ ├── subscription.yaml
│ └── operatorgroup.yaml
├── jobset/
│ └── kustomization.yaml
├── kueue-operator/
│ ├── kustomization.yaml
│ ├── namespace.yaml
│ ├── subscription.yaml
│ └── operatorgroup.yaml
├── leader-worker-set/
│ ├── kustomization.yaml
│ ├── namespace.yaml
│ ├── subscription.yaml
│ └── operatorgroup.yaml
├── opentelemetry-product/
│ ├── kustomization.yaml
│ ├── namespace.yaml
│ ├── subscription.yaml
│ └── operatorgroup.yaml
└── tempo-product/
├── kustomization.yaml
├── namespace.yaml
├── subscription.yaml
└── operatorgroup.yaml
Contains all infrastructure dependencies required for OpenShift AI.
Cluster-level configurations:
- default-kueue.yaml: Default Kueue configuration for cluster-wide job queuing
OpenShift operators managed via OLM (Operator Lifecycle Manager) with label-based categorization:
Monitoring Operators (monitoring=true):
- cluster-observability-operator: Cluster observability and monitoring
- tempo-product: Distributed tracing backend
- opentelemetry-product: OpenTelemetry observability
Serving Operators (serving=true):
- rhcl-operator: Red Hat Container Library (also auth)
- leader-worker-set: Leader-worker set management (also distributed workload)
- cert-manager: Certificate management (also distributed workload)
Distributed Workload Operators (distributedworkload=true):
- kueue-operator: Job queuing and resource management
- leader-worker-set: Leader-worker set management (also serving)
- cert-manager: Certificate management (also serving)
See dependencies/operators/README-labels.md for detailed usage instructions.
This repository contains operator configurations that can be deployed using standard Kubernetes tools:
# Apply all operators
kubectl apply -k dependencies/operators/
# Apply cluster configurations
kubectl apply -k dependencies/cluster-config/
# Apply operators by category using labels
kustomize build dependencies/operators/ | kubectl apply -f - -l monitoring=true
kustomize build dependencies/operators/ | kubectl apply -f - -l serving=true
kustomize build dependencies/operators/ | kubectl apply -f - -l distributedworkload=trueSee dependencies/operators/README-labels.md for detailed label-based deployment instructions.