Many things require database servers.

You can easily run one database container for each service that needs it, but it might be useful to provide a highly available database installation for the entire cluster.

Options:

• Avoid the Crunchy Data operator, see https://news.ycombinator.com/item?id=31883847
• Use the Percona Postgres operator
  • Use a solution like this to create users: https://www.percona.com/blog/manage-mysql-users-with-kubernetes/
• Use CloudNativePG https://cloudnative-pg.io/documentation/1.18/

If I do this, I would call it datadump.