2 Requirements #
This section describes the minimum platform prerequisites and recommended production sizing for SUSE Private Registry.
2.1 Prerequisites #
A Kubernetes cluster in version 1.33, 1.34 and 1.35
Helm version 3.2.0 or higher
Persistent Volume (PV) provisioner support in your infrastructure
An active subscription for SUSE Private Registry
2.1.1 Which Kubernetes versions are supported? #
Private Registry versions 1.0, 1.1 and 1.2 are currently validated on Kubernetes versions 1.33, 1.34 and 1.35.
| Private Registry version | Kubernetes versions |
|---|---|
1.0 | 1.33, 1.34, 1.35 |
1.1 | 1.33, 1.34, 1.35 |
1.2 | 1.33, 1.34, 1.35 |
Versions outside 1.33 through 1.35 are not currently validated for this release.
2.2 Hardware and sizing recommendations #
Use these values as a production starting point. Tune based on retention policy, image churn, scan concurrency and replication traffic. ==== Cluster Baseline
| Scope | Recommended starting point |
|---|---|
Worker nodes | 3 worker nodes minimum |
Node shape | 4 vCPU and 16 GiB RAM per node minimum |
Preferred node shape | 8 vCPU and 32 GiB RAM per node for higher scan and push concurrency |
These recommendations align with Rancher guidance for RKE2 Kubernetes. For details, see RKE2 Kubernetes installation requirements.
2.2.1 Persistent Storage Baseline #
| Component | Default chart size | Recommended starting size |
|---|---|---|
Registry data | 5 Gi | 500 Gi to 1 Ti |
Trivy cache and DB | 5 Gi | 20 Gi to 50 Gi |
Jobservice logs | 1 Gi | 10 Gi |
Internal PostgreSQL (if used) | 1 Gi | 20 Gi minimum |
Internal Valkey (if used) | 1 Gi | 20 Gi minimum |
The chart defaults are installation-oriented and should not be used as long-term production capacity values.