EDB Gives CloudNativePG Klio Backups and Previews Kubernetes Enterprise Tools
EDB is adding a more continuous recovery path to CloudNativePG while preparing paid enterprise features aimed at keeping teams on the open source operator. The announcements were made at KubeCon + CloudNativeCon North America 2026 in Salt Lake City, a day after CloudNativePG entered the CNCF Incubating program on Nov. 10. The timing links two different questions facing users of a growing open source project: how to limit data loss when a database fails, and how to obtain enterprise support without replacing the software teams already run.
The backup tool, Klio, is being donated to the CloudNativePG community. EDB is also previewing a supported Klio distribution and an EDB Postgres AI Plugin that adds enterprise capabilities, including transparent data encryption, while leaving the operator in place. Both offerings are planned for general availability by the end of 2026, and community images will remain free. That gives users a prospective choice between community software and paid support, though the announcements alone don't establish how the enterprise features will work in practice or what adopting them will require.
Continuous WAL streaming changes recovery trade-offs
CloudNativePG’s built-in backup process archives PostgreSQL write-ahead log segments after they’re complete. EDB says that approach can leave up to five minutes of transactions exposed during an outage. Klio instead streams WAL continuously and keeps a local copy beside the database cluster, a design intended to narrow the gap between committed activity and recoverable data. It can also let restores run at disk speed rather than requiring a full copy to be fetched from remote storage.
The distinction matters because backup speed and recovery speed aren't the same problem. A system may have copies of its data and still take time to restore service, particularly when recovery depends on retrieving information from remote storage. Klio’s local copy is meant to reduce that delay, while continuous streaming addresses the potential loss of recent transactions. The trade-off is operational: teams need to understand where local copies sit, how they’re protected, and how they fit with their existing storage plans. EDB describes the intended benefits, but the announcement doesn't provide comparative recovery results.
Klio encrypts data in transit and at rest, using customer-controlled keys that can be rotated. It can write to local volumes, S3-compatible storage, or both, giving operators options for where backup data lives. A single Klio server can protect multiple clusters and provide a shared view of backup health. That combination is relevant to teams running several Kubernetes databases, where separate backup routines can make it harder to see whether recovery coverage is current. A shared view may help, but its value will depend on how clearly it reports failures and how easily teams can act on them.
This is more significant than it looks: recovery tools are judged during an incident, but their usefulness depends on routine operational visibility beforehand. (And this is the part most people overlook.) A backup that exists but can't be located or restored in the time a team needs is of limited help. The announcement points to monitoring and storage flexibility as part of Klio’s proposition, though users will need to assess those capabilities against their own recovery procedures.
EDB extends its Kubernetes data plans
EDB’s enterprise Klio distribution is planned to include certified images, support, and long-term-support releases. These are familiar concerns for organizations that need a defined maintenance path and an accountable support channel alongside open source software. The company hasn't said here how those offerings will be priced or how their service terms will compare with community use, so the distinction currently rests on the announced support and release model.
The separate Postgres AI Plugin is intended to add enterprise features to CloudNativePG while leaving the operator in place. That matters for teams that have built deployment and operational processes around the operator: adding capabilities without changing that foundation could reduce the disruption of adopting them. But the practical test will be compatibility and day-to-day management. The plugin is still being previewed, and the announcement doesn't detail its full feature set beyond citing transparent data encryption as an example.
EDB is also tying the Kubernetes announcements to its recent acquisition of Altinity, which makes the Altinity Kubernetes Operator for ClickHouse. The company is pitching Postgres and ClickHouse as a multi-model data layer on Kubernetes: Postgres for transactions and ClickHouse for analytics, with change data capture feeding the latter. The shared declarative operating model is meant to give teams a common way to manage both database systems. That's an appealing direction for organizations already operating both kinds of workloads, although a common operating model doesn't by itself remove the work of coordinating data flows or diagnosing problems across systems.
Integrated enterprise ClickHouse capabilities are planned for the fourth quarter of 2026. As with the Klio distribution and Postgres AI Plugin, that timing sets an expectation rather than demonstrating a finished product. EDB is outlining a broader portfolio around Kubernetes, spanning database operations, recovery, and enterprise support. The breadth may suit teams looking to standardize how they run data services, but the offer will ultimately depend on whether the pieces work together in ordinary operations, not simply whether they share a platform story.
Open source adoption will test the model
EDB CTO Max Romanenko said the company aims to support the open source communities behind Postgres, ClickHouse, and cloud-native software, while giving enterprises a foundation they can run on infrastructure they control. The two goals can align when paid products fund support and development around community software. They can also create questions for users about which capabilities remain available in community editions and how closely the paid versions track upstream projects. The announcement says community images will remain free, but doesn't resolve every detail of that relationship.
If you're working in this space, the immediate questions are practical: whether continuous WAL streaming fits your recovery targets, whether local copies suit your storage design, and whether the proposed enterprise support addresses gaps in your operations. The numbers here are underwhelming in one respect: the announcements offer a five-minute exposure figure for the existing backup process, but no comparative recovery measurements for Klio. EDB’s plans could make CloudNativePG more attractive to organizations that need managed support and additional controls. The near-term test is whether Klio’s streaming backups and the enterprise add-ons fit naturally into CloudNativePG operations as adoption grows.
