Installing inside an AZ
In this model, the management cluster runs on the same Kubernetes cluster as one of your Superphenix availability zones. The operator, web console, and GitOps components live alongside the workload stack on that AZ.
Part of the deployment guide. For the alternative placement, see Installing outside an AZ.
Overview
The operator and console run on one of the workload clusters. This is easier to bootstrap but creates a dependency on the hosting AZ.
When to choose this model
- Single AZ or isolated AZs: each AZ has its own management cluster and control plane manages only that AZ.
- Security boundaries: strict isolation between AZs, with no shared management plane.
- Labs and first deployments: simplest path when you start from one hyperconverged cluster.
Requirements
- The management cluster must be a pre-existing Talos Linux cluster before you install the
superphenix-operatorHelm chart. - Bootstrap that cluster with Manual OS installation (or an existing Talos cluster you already operate).
- Size the AZ for both workload and management components on the same nodes.
Trade-offs
| Aspect | Notes |
|---|---|
| Deployment | Easier to deploy: one cluster to provision |
| Maintenance | Easier day-to-day: no separate management fleet |
| Multiple AZs | If one management cluster controls several AZs, the hosting AZ is a single point of failure |
| Failover | Management is tied to one AZ; failover is harder |
| Dependency | Management runs on Superphenix (same stack as the AZ) |
How to install the operator
Installation of the operator is typically done via Helm on the AZ cluster that will host management:
helm upgrade --install superphenix-operator \
ghcr.io/super-phenix/charts/superphenix-operator \
--namespace superphenix-system \
--create-namespace
When the operator runs on the same cluster as the AZ, define that cluster with connection.mode: Local. The operator connects in-process and does not need a remote API URL or credentials secret.
apiVersion: operator.superphenix.net/v1alpha1
kind: Cluster
metadata:
name: cluster-local
namespace: superphenix-system
spec:
deploymentTopology: Hyperconverged
region: us-west-2
availabilityZone: us-west-2a
version: 0.0.0-latest
connection:
mode: Local
systemConfiguration:
apps: {}
See Configure a cluster for the full Cluster resource reference.
Next steps
- Manual OS installation: bootstrap Talos on the AZ that will host management (if not already done).
- Install the operator using the Helm command above.
- Installing an AZ: define
Clusterresources for your AZs (Localfor the hosting cluster,Remotefor any others).
See Deployment topology for the full comparison with management outside an AZ.