Supabase Kubernetes

There are two primary ways to run Supabase on a Kubernetes cluster:

1. Supabase Kubernetes Operator: A Kubernetes-native approach using Custom Resources (core.supabase.io/v1alpha1). This allows you to manage Supabase through modular resources. Note that the Operator is in early development and its API may change.
2. Supabase Helm Chart: A traditional deployment method using Helm charts. Detailed information can be found in the charts/supabase directory.

The Operator manages Supabase via several Kubernetes Custom Resources (CRDs). This modular design allows you to enable only the specific components your project requires.

Supported Custom Resources:

• Project (projects): Represents a complete Supabase instance.
• SingleDatabase (singledatabases): A Postgres database managed by the Operator.
• Function (functions): Edge Functions deployed within the cluster.
• Migration (migrations): Used to apply SQL scripts to databases or manage internal Supabase upgrade migrations.

Use the supabase-project chart to create a SingleDatabase and a Project resource. You must provide specific configuration values for hostnames, ports, and authentication URLs. The Operator will then automatically provision the Postgres StatefulSet, component Deployments, Services, Secrets, and run sync Jobs for JWT and database configuration.

helm install supabase supabase/supabase-project \
  --set fullnameOverride=supabase \
  --set project.http.hostname=localhost \
  --set project.http.port=8000 \
  --set auth.siteUrl=http://localhost:3000 \
  --set studio.orgName="Default Organization" \
  --set studio.projName="Default Project"

Install the Operator into the supabase-operator namespace. This command installs the necessary Custom Resource Definitions (CRDs) and deploys the controller in a single step.

helm install supabase-operator supabase/supabase-operator \
  --namespace supabase-operator \
  --create-namespace

To remove the local Kind cluster used for development, run the following command:

make kind-down

To run the Supabase Kubernetes Operator locally for development, use a Kind cluster. This sequence sets up the cluster, generates the necessary code and manifests, installs the CRDs, and starts the controller.

Note: Ensure you have kubectl configured and Kind installed before starting.

make kind-up
make generate
make manifests
make install
make run
kubectl apply -k config/samples

To access the Supabase Studio and APIs locally, forward the supabase-envoy service to your machine. Once forwarded, the services will be available at http://localhost:8000.

kubectl port-forward svc/supabase-envoy 8000:8000

You can automate database schema migrations by adding them to the db.config field. Any SQL scripts provided here will be applied during the database initialization process. Use the multi-line YAML syntax (|) to include the SQL content.

db:
  config:
    20230101000000_profiles.sql: |
      create table profiles (
        id uuid references auth.users not null,
        username text unique
      );

Before deploying, add the official Supabase Helm repository to your local Helm configuration and update your local charts list.

helm repo add supabase https://supabase-community.github.io/supabase-kubernetes
helm repo update

Starting with chart version 0.7.1, environment.<component> is an array of Kubernetes env entries. Each entry supports value: or valueFrom:.

User-provided entries are rendered after the chart defaults, allowing you to override any chart-managed environment variable by providing an entry with the same name.

environment:
  auth:
    - name: GOTRUE_API_HOST
      value: "0.0.0.0"
    - name: GOTRUE_EXTERNAL_AZURE_SECRET
      valueFrom:
        secretKeyRef:
          name: azure-secret
          key: secret