new-application-team.md

Adding a new application team

Application teams are empowered to deploy, run, and maintain their own software on top of the platform, in a self-service fashion. The efforts that follow allow to:

• isolate different teams
• make them more autonomous
• lower their cognitive load

Checklist

1. Decide on a reserved Kubernetes namespace e.g. sciety.
2. Create a Github repository that will contain definitions for all Kubernetes resources. Optionally create a deploy key and add to this repository in github.
3. Sync this repository by adding a Kustomization for the team, and a Flux Kustomization.
4. In the kustomizations, create (as needed):
   1. Create a dedicated node pool that will host all workloads in a certain namespace.
   2. Create a dedicated storage class that will configure and tag storage for a namespace.
   3. Create a deployment sync object for the git repository created above, referencing the deploy key secret if created above
5. Test the setup via the creation of the desired Kubernetes namespace. Note: Make sure to set the annotation to assign workloads to the correct nodepool:

   apiVersion: v1
   kind: Namespace
   metadata:
   name: application-team-1
   annotations:
       elifesciences.org/default-project: application-team-1