EDB Adds Continuous WAL Protection and Enterprise Features to CloudNativePG

Nov 11, 2026 327 views

CloudNativePG is gaining a continuous backup option and a path to enterprise features without changing its Kubernetes operator. At KubeCon + CloudNativeCon North America 2026 in Salt Lake City, EnterpriseDB (EDB) donated its Klio backup and recovery tool to the community and previewed two commercial offerings built around CloudNativePG (CNPG). The approach gives users a choice: take the community tool as a starting point, or buy packaged support and additional controls while continuing to use the same operator. That distinction matters to organizations wary of adopting a product path that could require a change to their database platform.

The announcements followed CNCF’s acceptance of CNPG as an Incubating project on Nov. 10. They also came less than a week after EDB completed its acquisition of Altinity, which develops the Altinity Kubernetes Operator for ClickHouse. Together, the news points to EDB’s interest in building a broader set of data services around Kubernetes. The announcements themselves, however, don’t establish how tightly those services will be integrated in practice or how much work operators will need to do to run them together.

Continuous WAL streaming narrows recovery gaps

CNPG’s built-in backup options archive Postgres write-ahead log (WAL) data in completed segments. According to EDB, that can leave as much as five minutes of transactions at risk during an outage. Klio streams WAL continuously and retains a local copy alongside the cluster, aiming to reduce data loss and let restores run at disk speed instead of retrieving a full copy from remote storage. The difference is a matter of how frequently recovery data becomes available: segment-based archiving waits for a segment to complete, while continuous streaming is designed to capture changes as they happen.

That can matter when a database supports applications that keep writing data throughout the day. A recovery point further behind the failure can mean more transactions need to be reconstructed or entered again. But the stated benefit is an aim, not a guarantee that every restore will be fast or that no data will be lost. Results will depend on the system’s storage, configuration, and failure conditions, details EDB hasn’t supplied in this announcement.

Klio’s local copy is central to its speed claim. Restoring from data already held near the cluster can avoid waiting for a full backup to travel from remote storage, though local retention also makes capacity and storage management part of the operational picture. The design combines that local copy with the option to store backups on local volumes, S3-compatible storage, or both. That flexibility gives administrators room to fit backup placement to existing infrastructure, but it also leaves them responsible for choosing and maintaining a setup that meets their recovery needs.

Klio encrypts data in transit and at rest using customer-controlled keys that can be rotated. It supports local volumes, S3-compatible storage, or both. A single server can protect multiple clusters, with backup health visible in one place. That shared view could help teams spot backup problems across a fleet without checking each cluster separately. It also creates a wider operational dependency: administrators will need to understand how the server, keys, storage targets, and protected clusters fit together.

Paid features keep the community operator

EDB is previewing a supported distribution of Klio with certified images and long-term-support releases. A separate EDB Postgres AI Plugin for CNPG is set to add enterprise capabilities, including transparent data encryption, while retaining the open-source operator. EDB plans general availability for both offerings by the end of 2026; community images will remain free. For teams already using CNPG, the proposed model could let them add support and selected features without replacing the operator at the center of their deployment.

There’s a meaningful distinction between keeping an operator open source and making every related capability freely available. A paid distribution can offer organizations a clearer support channel and a longer maintenance window, while the community version remains available for users who don’t need those services. The details will determine how useful that split is: the announcement doesn’t specify pricing, the scope of support, or how the commercial plugin’s capabilities will be packaged.

“By continuing to invest and give back across the open source communities behind Postgres, ClickHouse software, and the cloud-native ecosystem, we give enterprises the foundation to build that future at scale, on infrastructure they own and control,” said Max Romanenko, CTO of EDB.

The promise of keeping customer infrastructure under the customer’s control is familiar in cloud-native software, but it’s not the whole procurement question. Enterprises also need to weigh maintenance, support, and the effort required to operate the resulting stack. This is more significant than it looks: a commercial layer that preserves an existing operator may lower the friction of adoption, but its value will depend on what it adds beyond a label and a support contract.

Postgres and ClickHouse share an operating model

EDB is also pitching Postgres and ClickHouse as complementary parts of EDB Postgres AI, running natively on Kubernetes. The company says teams can use the operators’ shared declarative approach to manage transactional Postgres alongside ClickHouse analytics fed by change data capture. In that arrangement, Postgres handles transactional workloads while changes are passed along for analysis in ClickHouse. Using Kubernetes operators for both can give teams a familiar way to describe and manage services, even though the databases serve different roles.

That shared operating model doesn’t automatically make the data flow simple. Change data capture has to move changes reliably, and teams still need to monitor the path from the source database into analytics. If you’re working in this space, the practical question is whether the operators make those tasks easier to coordinate, not simply whether both databases can run on Kubernetes. (And this is the part most people overlook.)

EDB plans to introduce integrated enterprise ClickHouse capabilities in the fourth quarter of 2026. The announcement provides a direction and a target, but not enough detail to assess how the integration will work or what capabilities it will include. For now, the company’s pitch rests on bringing database operations under a common Kubernetes pattern and pairing transactional data with analytics.

The operational test comes next

For teams standardizing data services on Kubernetes, the next test will be whether continuous recovery, enterprise controls and analytics operations can fit into one practical workflow. Klio’s continuous WAL streaming addresses a specific recovery limitation described by EDB, while the planned commercial offerings aim to add support and enterprise features around the existing operator. The ClickHouse plan extends the same operational story to analytics, but remains a future product announcement.

What this means for you depends on the gap you’re trying to close. If recovery time and data loss are the immediate concerns, Klio’s local copy and continuous stream are the claims to examine first. If the larger aim is to run several data services through Kubernetes, the shared operator model may be more relevant. Either way, the announcement sets expectations; deployment experience will show whether the pieces deliver a workflow teams can manage day to day. For now, the direction is clear. The proof is still ahead.

Source: Techstrong Editorial · cloudnativenow.com

Comments

Sign in to comment.
No comments yet. Be the first to comment.

Related Articles

EDB Donates Klio Backup Tool to CloudNativePG, Previews E...