Skip to content

Latest commit

 

History

History
51 lines (38 loc) · 2.67 KB

SCENARIOS.md

File metadata and controls

51 lines (38 loc) · 2.67 KB
Raw

Deployment scenarios with Kubernetes

If you don't see your scenario in this document, we would welcome a pull request.

We identified different scenarios for deploying Plone with Kubernetes. This is not meant to be complete, but should work for 80% of the use cases. Other use cases can build on top of these scenarios.

The scenarious are written from simple to complex and build on top of each other.

Scenario 1: Plone Classic UI using PostgreSQL

Plone Classic UI with PostgreSQL is the simplest scenario. It uses a database, stores the blobs in the database, and uses Classic UI, which is provided by the plone-backend container. The image is used as-is, without any additional add-ons.

Infrastructure

plone-backend will be provided as a Deployment with at least two replicas. A LoadBalancer service bundles the plone-backend pods.

The database is provided by a PostgreSQL StatefulSet with one replica. It needs a PersistentVolumeClaim to store the data.

The plone-backend pods connect to the PostgreSQL pod.

Scenario 2: Plone with Volto using PostgreSQL and Routing capable Ingress

Plone with Volto using PostgreSQL is the next step. Volto is the default Plone 6 frontend. This scenario is the minimal default setup for a modern Plone 6 site in Kubernetes with a routing-capable ingress controller. It uses a PostgreSQL database and stores the blobs in the database. It provides the Plone backend API server, which is run by the plone-backend image. The Volto frontend is provided by the plone-frontend image. Both images are used as-is, without any additional add-ons.

Infrastructure

Additional to scenario 1, this scenario needs a plone-frontend pod. plone-frontend will be provided as a Deployment with at least two replicas. A LoadBalancer service bundles the plone-frontend pods.

Also, it needs an ingress with HTTP routing capabilities, such as a Traefik ingress controller. Kubernetes environments without such an ingress controller are covered by scenario 3. The ingress controller will be configured to route the requests to the plone-frontend and plone-backend pods, dependent on the requests path.

Scenario 2: Plone with Volto using PostgreSQL and Non-Routing capable Ingress

Plone with Volto using PostgreSQL and Non-Routing capable Ingress is the same as scenario 2, but without the need for a routing-capable ingress controller.

Infrastructure

Additional to scenario 2 this scenario needs an nginx pod. The nginx pod will be configured to route the requests to the plone-frontend and plone-backend pods, depending on the requests path. nginx will be provided as a Deployment with at least two replicas. A LoadBalancer service bundles the nginx pods.