In its ongoing efforts to build out a full stack of open source cloud native technologies, the Cloud Native Computing Foundation has accepted an open source Kubernetes operator for the PostgreSQL relational database system, called CloudNativePG (CNPG), into its sandbox of experimental technologies.
PostgreSQL enterprise support provider EnterpriseDB (“EDB”) created the open source operator two years ago and submitted it to the CNCF in 2024.
“Kubernetes manages the availability of the different applications and container applications; we’re able to build our operator to key off of that,” said Jozef de Vries, chief product engineering officer at EDB, in an interview with TNS.
The CNCF sandbox program provides an early incubation stage for open source cloud native technologies, so they can gain visibility and work more seamlessly with other CNCF cloud native technologies.
CloudNativePG is one of a number of operators to help PostGreSQL fit into Kubernetes environments, alongside Crunchy Data Postgres Operator, Zalando, KubeDB and StackGres. EnterpriseDB claims CloudNativePG is the most popular of this lot, given its 5,000 GitHub stars from individuals who have downloaded the open source code.
“This milestone is a testament to the transformative power of Kubernetes in revolutionizing enterprise data management in the AI era,” said Stefanie Chiras, Red Hat senior vice president, in a statement. “EDB Postgres AI, running on Red Hat OpenShift AI, can empower organizations to more easily deploy scalable, high-performing databases while supporting and accelerating AI-ready applications.”
No More Database Tools Kubernetes Operator
A Kubernetes operator provides a way for Kubernetes to manage the deployment of complex applications, so they do not need to be set up manually each time they are deployed. Databases have been particularly difficult for Kubernetes, a resource orchestrator for distributed applications, due to their need for permanent storage.
The initial use case for Kubernetes was for stateless applications, or apps that did not rely on permanent data storage to operate. Databases, whose job is to hold permanently data, are the definition of “state.”
Operators such as CloudNativePG have proved to be crucial at tying the two worlds together.
Setting up a traditional database system, “particularly when you have requirements for high availability and backups and change management, you often have to then incorporate a lot of additional tooling and other technologies to ensure all those sorts of use cases are fulfilled, and you end up with a fairly complicated set of of technologies all to make sure your database systems are doing what you need to do,” de Vries said.
“Where we’ve benefited from Kubernetes, as it relates to Postgres in particular, is that we’ve been able to orchestrate the high availability of the database system deployment within Kubernetes itself,” de Vries said. The HA and failover capabilities are built into the operator itself.
With CloudNativePG, Kubernetes has instructions about, for instance, where the primary copy of the database is, where the secondary copy is located, and when it should fail over from one to the other, de Vries explained.
One EDB customer noted on G2 that it was able to stand up replicas of PostgreSQL over multiple Kubernetes clusters, so the database can be deployed up or down depending on the amount of traffic it is receiving at that time.
The post Kubernetes Operator for PostgreSQL Hits the CNCF Sandbox appeared first on The New Stack.
