\

Bart Farrell: In this episode of The Landscape, I speak to Gabriele Bartolini, who is leading all things PostgreSQL at EDB and is the maintainer of CloudNativePG—a PostgreSQL operator for Kubernetes that is currently applying to become a CNCF project. Gabriele sheds light on how the process is going and what is required for a project to join the CNCF. Gabriele, a fantastic human being, also comes from the beautiful Italian city of Prato.

This podcast is brought to you by OVHcloud, the global cloud provider delivering industry-leading performance and cost-effective solutions. From managed Kubernetes to private registries, OVHcloud has the cloud-native tools you need. OVHcloud: the world needs different.

Now let’s check out the episode with Gabriele.

---

Gabriele Bartolini: My name is Gabriele Bartolini. I’m VP and Chief Architect of CloudNative at EDB. EDB is a company that develops and promotes the usage of PostgreSQL everywhere, including in Kubernetes.

Bart: Wonderful. Now, what project are you currently maintaining?

Gabriele: My main project right now is CloudNativePG. CloudNativePG is a Kubernetes-native operator that allows you to run PostgreSQL databases in a fully declarative way inside Kubernetes.

Bart: What is this project solving? What are people struggling with that made it necessary to develop CloudNativePG?

Gabriele: PostgreSQL is the most used database out there. The trend we’ve seen with deployments on bare metal, VMs, and now managed services is becoming equally popular in Kubernetes. We’re seeing a lot of developers and infrastructure teams who want to move PostgreSQL into Kubernetes.

Our operator provides a new perspective on managing PostgreSQL, offering a fully open-source, vendor-neutral, community-driven project.

Bart: What alternatives currently exist, and how is your project different?

Gabriele: Of course, you can run PostgreSQL outside of Kubernetes. A popular use case is running PostgreSQL as a database-as-a-service in a traditional way on VMs. Inside Kubernetes, there are a few options.

One option is to run PostgreSQL without an operator, which I highly discourage. There are other operators as well, such as the ones from Crunchy Data and Zalando, which have been widely used. However, CloudNativePG is quickly becoming the first choice, especially as we’ve applied for CNCF Sandbox status and are hopeful for acceptance in the next round.

Bart: For people unfamiliar with the process, what’s it like to apply to become a CNCF Sandbox project? What’s required—end users, GitHub stars, contributors?

Gabriele: First, we need adoption. Over the past few years, we’ve worked to increase adoption of the operator. While we’ve looked at GitHub stars, they don’t mean much. For us, what matters is organic growth. We’ve avoided marketing campaigns to ask for stars, focusing instead on working with existing EDB customers and the community to grow adoption.

We’ve engaged end users like IBM Cloud Pak, Google with GKE, and Azure with AKS, all of whom use our operator. Startups like TempoDB, which provides database-as-a-service, have also adopted it.

We’ve worked with the CNCF Storage TAG (Technical Advisory Group) to contribute ideas and use cases back to Kubernetes storage. For example, one of our contributors is working on Kubernetes storage, including features like volume snapshots. The three key factors are adoption, contributors, and collaboration with CNCF TAGs relevant to the project.

Bart: In terms of the next steps for CloudNativePG, what can we expect in the coming months?

Gabriele: We’re working on making the operator pluggable. Right now, it’s difficult to introduce new features, so we’re simplifying and refactoring the codebase to include hooks. We’re starting with backup and recovery tools, enabling third parties to write plugins.

We’ll also expand observability and continue adding new features, such as support for declarative databases and logical replication. One exciting feature we’re working on is enabling zero-downtime migration from services like AWS RDS into Kubernetes using CloudNativePG.

Bart: What’s your favorite part about being a maintainer?

Gabriele: The ability to interact with end users. Being part of the community and receiving their feedback creates a short feedback loop that helps us improve the project. I’ve been in IT and open source for many years, but the momentum CloudNativePG has right now is incredible. It’s a privilege to be part of this phase.

Bart: How can people get in touch with you?

Gabriele: The best way to reach me is on LinkedIn.

Bart: Is there anyone in the project you’d like to give a shout-out to?

Gabriele: Absolutely. Leonardo Cecchi and Michele Querci are the brains behind the project—I’m more of a spokesperson now. Other maintainers include Jonathan Gonzalez, Francesco Canovi, Armando Ruocco, and Philippe Scorsolini, who is also a Crossplane maintainer. Finally, a big thank you to all our users, who inspire us to keep working hard.

Bart: AI is a big topic right now. Does CloudNativePG have any AI-related features?

Gabriele: Yes. CloudNativePG includes PGVector, a popular PostgreSQL extension that allows it to act as a vector database. Our job is to provide a reliable framework to store data for AI use cases. We stop at the database layer, ensuring it’s reliable and scalable, while AI applications can run on top of it seamlessly.

Bart: Thank you very much for your time today, Gabriele!

Gabriele: Thank you! Cheers!