Set up Metabase in a Kubernetes cluster with a Postgres backend.
If you want to stand up an example of Metabase locally you can do so just after you install:
- kind - Uses Docker to create a Kubernetes control plane and worker nodes on your local machine.
- kubectl - The Kubernetes command line tool.
After you've installed these two tools, you can stand up a properly configured kind cluster by running:
bash local/start_kind.sh
This command creates a local kind cluster with HTTP ingress (see local/kind_config.yml for more details), using Contour (see local/start_kind.sh for more details).
After the kind cluster is up and running, you can simple execute:
kubectl apply -f manifests/local
This will create environment variables for the Postgres and Metabase containers, expose Postgres and Metabase as services, and make Metabase available to you at 127.0.0.1. You can check the status of your pods with kubectl get pods.
Metabase comes with a sample data set out of the box, so you can skip configuring your own data sources if you so choose.
When you are done with your cluster, you can run the following to delete your resources:
kubectl delete -f manifests/local
Finally, you can delete your cluster with:
kind delete cluster
The local manifests are set up to be easy to run, and you should to be able to point at a remote k8s control plane (that is configured for HTTP ingress) and achieve the same results, however you may want to create a PersistentVolume for the Postgres database to ensure your Postgres backend is truly persistent.
Metabase is very easy to deploy, all you need to do is point a Metabase container at a Postgres database and then start using the Metabase GUI! It very much horizontally scalable, since every Metabase container simply needs to be pointed at the Postgres backend.
Additionally, Metabase has a full REST API, which can be very useful for programmatically setting up user access, database hooks, table documentation, and more.