Introduction & Deployment Approach
The OLIVE Platform is a cloud-agnostic Kubernetes application designed to follow Cloud Native Computing Foundation (CNCF) practices. You can deploy it on public clouds, private datacenters, or edge infrastructure while keeping a consistent operational model.
We follow the CNCF-recommended setup procedure for Kubernetes workloads, using Helm charts distributed through an OCI registry.
The platform is tested with the Kubernetes distribution RKE2—using a different Kubernetes distribution might require additional steps.
This document describes the chart registry, required access, and the two supported deployment variants.
Chart & Image Registries
- Helm chart registry (OCI):
oci://registry.gitlab.com/olive6224315/helm-charts/charts - Image registry (OCI): Provided within the Helm chart values; images are published to our GitLab registry.
For the full OLIVE delivery package (GitLab projects, Helm and container OCI endpoints, GitOps blueprint, and scenario examples), see Registries and Repositories.
Access to the Helm chart registry requires an access token. Please request this from the OLIVE team. Contact: olive-support@boc-group.com.
Once you have it, test it with the hello-world chart before deploying anything real — fetching and rendering the chart is enough and touches no cluster. See Verify registry access.
Deployment Variants
We support two modes of deployment, both using the same Helm chart with different values:
- Development deployment: simplified, single-tenant setup without authentication.
- Production deployment: multi-tenant, authentication-enabled setup split into stateless resources (installed once) and per-tenant resources.
Those variants are product modes; choosing a deployment location and tenant topology is covered in Deployment Options and Tenant Isolation.
Assumptions
This documentation is designed for Platform Administrators. It does not define the role as a job: what follows is what this guide assumes you can already do, because OLIVE is built on these tools and the documentation does not re-teach them.
- Kubernetes — namespaces, RBAC, and reading a failing pod
- Helm — values files, and debugging what a chart actually rendered
- Argo CD — Applications, sync waves, and why a sync stalls. The OLIVE charts declare their own sync waves and Argo CD has to honour them, so this is the one worth being comfortable with
What you will and will not be able to do yourself
OLIVE deploys into namespaces you own. Several things it depends on are cluster-scoped, and if you do not hold cluster rights you will be requesting them rather than applying them: CRDs and admission webhooks, ClusterRoles, the mesh configuration, and the cluster-scoped Kyverno policies.
Infrastructure Prerequisites is written for that handover — it says what each component must provide, and which checks are yours to run versus your cluster operator's.
Dependent infrastructure technologies:
The following infrastructure technologies must be pre-installed and configured in your Kubernetes cluster before deploying OLIVE. These are typically provisioned and managed by your cluster operator:
- HashiCorp Vault – Secrets management and encryption; OLIVE uses this for storing sensitive data and credentials
- Keycloak or equivalent OIDC provider – Identity and access management; provides authentication and authorization for OLIVE users
- Traefik Ingress Controller – Ingress and traffic routing; RKE2 provides Traefik by default. Traefik is the only supported ingress controller from 1.3 onwards
- CloudNativePG (CNPG) operator – PostgreSQL database operator; provides managed PostgreSQL instances for OLIVE Platform state
What each of these must provide before OLIVE can run on it is in Infrastructure Prerequisites; the full technology map is in Platform Technology Overview.
Deployment Methods
We provide setup instructions via three approaches:
- Install with Helm CLI – A manual installation path (⚠️ not fully tested; use with caution).
- Install with Argo CD – Our validated and recommended approach, making heavy use of sync waves to control the installation order of stateless and tenant resources.
- Install via Rancher – Register the OCI chart and install from the Rancher UI.
Next Steps
After understanding the deployment variants and methods, proceed to:
- Registries and Repositories – Understand the delivery package, access points, and available registries for Helm charts and container images.
- Deployment Options and Tenant Isolation – Decide where to deploy (lab, on-house, cloud) and choose your tenant isolation strategy (silo or multi-tenant).