Declarative Database & User Management in DatabaseCluster CR

Percona Everest
1

Hi Percona Team and Community,

First of all, I want to say a huge thank you to the contributors for Percona Everest. It is an amazing tool that has significantly simplified our database provisioning on Kubernetes.

I am writing to ask about the current state and roadmap for declarative user and database management within the Everest DatabaseCluster Custom Resource.

I have noticed that the underlying operators used by Everest already have support for defining users and databases directly in their manifests under the spec.users options (as done here in the PXC operator).

Currently, when using the DatabaseCluster CR in Everest, I haven’t found a documented way to pass these user/database definitions down to the underlying clusters, and I have to create them by imperatively interacting with the cluster.

  1. Does this feature already exist? Is there a section in the DatabaseCluster spec (perhaps a pass-through or a specific users field) that I might have missed in the documentation?

  2. If not, is it on the roadmap? With the individual operators now supporting this “GitOps-friendly” approach to initializing schemas and application users, are there plans to expose this capability through the Everest operator in an upcoming release?

Being able to define the initial application database and user in the DatabaseCluster manifest would allow us to be fully declarative. Currently, we have to provision the cluster via Everest and then run separate out-of-band scripts (or manual SQL commands) to set up the actual application schemas and credentials.

Thank you again for your hard work on this project!

1 Like
2

Hi @eneriz,

Thanks for the feedback, messages like this one really helps us shape the product.

This is definitely in our roadmap but there isn’t any ETA yet.

1 Like