Installation

The operator ships three install methods. Pick the one that fits your delivery model:

Method	Best for
Helm chart	Most clusters; GitOps and templated installs
kustomize (config/)	Plain kubectl apply -k, no Helm
OLM bundle	OpenShift / OperatorHub.io and Operator Lifecycle Manager

All three install the same thing: the operator Deployment, its CRDs, RBAC, and (optionally) webhooks. The operator then installs upstream capability operators on demand from charts embedded in its binary.

Cluster-admin for the install

The operator manages CRDs, cluster RBAC, and webhooks across namespaces, so the install needs cluster-admin. Day-to-day use of the custom resources does not.

Option 1 — Helm

helm install navique-ai-core-operator \
  oci://ghcr.io/scigility/charts/navique-ai-core-operator \
  --namespace navique-system --create-namespace

Useful values (see the chart's values.yaml for the full list):

image:
  repository: ghcr.io/scigility/navique-ai-core-operator
  tag: ""               # defaults to the chart appVersion
# Restrict which namespaces the operator watches (empty = all namespaces).
watchNamespaces: []
resources:
  limits:   { memory: 1Gi }
  requests: { cpu: 100m, memory: 512Mi }
# Leader election is on by default for HA-safe single-active reconciliation.
leaderElection: true

Option 2 — kustomize

From a checkout of the operator repository:

# Install the CRDs.
make install

# Deploy the manager (set IMG to your image).
make deploy IMG=ghcr.io/scigility/navique-ai-core-operator:latest

Or apply the kustomize overlay directly:

kubectl apply -k config/default

CRD kustomization

When building from source, every CRD must be listed in config/crd/kustomization.yaml, or a kustomize install will silently omit it. The provided overlays already include all of them.

Option 3 — OLM / OperatorHub

On a cluster with the Operator Lifecycle Manager installed, build and deploy the bundle from source:

make bundle IMG=ghcr.io/scigility/navique-ai-core-operator:latest
make bundle-build bundle-push BUNDLE_IMG=ghcr.io/scigility/navique-ai-core-operator-bundle:latest
operator-sdk run bundle ghcr.io/scigility/navique-ai-core-operator-bundle:latest \
  --namespace navique-system

On OpenShift the operator also appears in OperatorHub once the catalog source is published.

Verify the install

# The manager should be Running.
kubectl -n navique-system get deploy navique-ai-core-operator

# All CRDs should be present.
kubectl get crds | grep core.navique.com

# The management console is deployed by default (no CR required).
kubectl -n navique-system get deploy,svc -l app.kubernetes.io/component=management-plane

You should see the operator Deployment available, the *.core.navique.com CRDs established, and the management-plane console running. The console deploys by default — the ManagementPlane resource only configures it (host, ingress, SSO).

What gets installed later, on demand

The operator does not install every upstream operator up front. Capability operators (CloudNativePG, External Secrets, cert-manager, …) are installed lazily — only when a custom resource needs them — and each is a standalone, provenance-aware Helm release. See Bundled Charts and Provenance & Lifecycle.

Next

Bring up a platform with the Quick Start, or tune the manager via the Configuration page.