Kubernetes made deploying easy. Nobody warned you about the databases.

The DevOps movement, set in motion nearly 20 years ago, broke down the traditional divide between software developers and IT operations and adopted the maxim of “you build it, you run it” as its own. However, in Kubernetes environments, that model can become unrealistic when application teams are also expected to operate databases and data services. In practice, DevOps database operations still require much more structure than many teams expect.

In my role as senior cloud platform engineer at the cloud-native services provider anynines, I’ve seen this time and time again: developers who are highly skilled and highly effective at building software have to deal with databases that sit in a completely different realm and require a different kind of operational expertise.

Developers who are highly skilled and highly effective at building software have to deal with databases that sit in a completely different realm

That’s especially true in platform engineering database management, where the goal should be to reduce operational burden, not shift it onto application teams. A strong internal developer platform should make platform engineering databases easier to consume, rather than harder to run.

Tools like the a9s Hub demonstrate this model in practice: by enabling developers to request databases through Kubernetes-native workflows while a centralized control plane handles provisioning, backups, patching, and governance.

It’s easy enough to deploy Postgres, Redis, or OpenSearch with a few steps. Run a Helm install command, for example, and the service is up and running, provided everything works. The challenge comes when something falters, and those services need to remain reliable, which means patching, monitoring, backing up, and properly recovering them. That’s not something most product teams were ever meant to build deep expertise in.

Modern applications rarely depend on a single data service. A software-as-a-service application might need Postgres on Kuberne…