Code FREAK Helm Chart

Official Helm chart for installing Code FREAK on a Kubernetes cluster.

Quickstart

helm repo add codefreak https://codefreak.github.io/charts
helm install my-codefreak-install codefreak/codefreak

Configuration values

Please check out the values.yaml for a list of supported values and their purposes.

Prerequisites

• Helm
• Kubernetes Cluster (version >=1.19 recommended for stable Ingress support)
• Database: Postgres recommended. By default, an in-memory database will be used. Data will get lost between restarts!

Resource requirements

The resource requirements will depend on the number of concurrent users you expect. For a production environment you should expect 1-2GiB RAM per concurrent user.

Creating the JWT signing keypair

Code FREAK will use signed JWTs for authenticating users. For this to work you have to create an RSA keypair once and store it in a secret that will be accessible from the backend application.

# Create a key pair and convert the private key to the PKCS8
openssl genrsa -out private.pem 2048
openssl rsa -in private.pem -pubout -out public.pem
openssl pkcs8 -topk8 -nocrypt -in private.pem -out private-pkcs8.pem
# Create a secret containing both values
kubectl create secret generic my-secret-name --from-file=private=private-pkcs8.pem --from-file=public=public.pem

Afterwards, set the following values in your values.yml:

workspaces:
  jwtKeypairSecretName: my-secret-name
  jwtKeypairSecretPrivateKey: private # default, can be omitted
  jwtKeypairSecretPublicKey: public # default, can be omitted

Workspace namespace

Code FREAK will start additional Pods on demand, called "workspaces". Workspaces are used to provide in-browser IDEs for students or evaluating their code. For security reasons we recommend starting workspace Pods in a dedicated namespace. To create a dedicated namespace you can set the following value in your values.yml:

workspaces:
  namespace: my-workspace-namespace

If your user does not have access to create namespaces please set workspaces.createNamespace to false and ask your cluster admin to create a dedicated namespace for you:

workspaces:
  # Must be created manually
  namespace: my-workspace-namespace
  createNamespace: false