I know there are several professional PostgreSQL Kubernetes Operators out there (crunchy, zalando, kubeDB, stackgres, percona).
But my use case is small and simple.
I thought about running PG in a sidecare together with my application image in a Kubernetes Pod.
All I need is a persistent directory.
It is no problem if the pod gets killed an recreated on a different node, as long as the mounted directory is preserved. I can live with a short downtime in this context.
Of course this means I can only run one pod. Scaling to several pods is not possible, but that's perfectly fine for my small use-case.
Unfortunately there are only articles about scaling PG up, but I found no article how to scale PG down (in the context of Kubernetes).
What is wrong with this approach?
