EDB Engineering Newsletter #17

Apr 30, 2026

For those paying attention: yes, we skipped a month. But that only means extra-bonus content this month! Scroll down for a look at what the EDB Engineering team has been up to. But first, some of the industry news headlines that grabbed our attention 📰

News we’re watching:

AI intensifies work

In an eight-month study of how generative AI changed work habits at a U.S.-based technology company with about 200 employees, it was found that employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day. While this may sound like a dream come true for leaders, the changes can be unsustainable, causing problems down the line, including cognitive fatigue, burnout, and weakened decision-making.

The researchers suggest companies need to develop a set of norms and standards around AI use—an “AI practice”: https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it

How Well Do Agentic Skills Work in the Wild: Benchmarking LLM Skill Usage in Realistic Settings

If you've invested in SKILL.md files for your agentic workflows, this paper is a reality check. Existing skill benchmarks are overly idealized and agents are given hand-crafted, task-specific skills that function more like solution guides than reusable knowledge. This study strips those safety nets away, requiring agents to retrieve skills from a collection of 34k real-world open-source skills, and the results are stark: Claude Opus 4.6 drops from 55.4% with force-loaded skills to 38.4% when retrieving its own, barely above the 35.4% no-skill baseline.

The good news: query-specific refinement partially recovers this loss, improving Claude's pass rate from 57.7% to 65.5% on Terminal-Bench 2.0. But only when the retrieved skills are reasonably relevant to begin with. The takeaway for anyone building skill infrastructure is that retrieval quality is the bottleneck, and more skills won't help if the retrieval pipeline is weak.

David Steele stops work on pgBackRest

David Steele announced that he stopped working on open source backup tool pgBackRest.

I have been maintaining pgBackRest and looking for a position that would allow me to continue the work, but so far I have not been successful. Likewise, my efforts to secure sponsorship have also fallen far short of what I need to make the project viable.

A challenge many maintainers, especially those without co-maintainers, are (maybe increasingly) struggling with: making a living, and investing time in a project that is used by many.

An end of an era to be sure, and we’re curious about the community’s response. Our Gabriele Bartolini shared his perspective on his blog:

When corporate sponsorship disappears and community sponsorship cannot fill the gap, maintainers face an impossible choice: do it poorly, or stop. David chose the latter. He is not the first, and he will not be the last.

From the EDB team:

Shaping SQL in São Paulo

EDB engineers Euler Taveira and Matheus Alcantara (to the far left on the picture) attended the ISO/IEC SQL Standards Committee meeting in São Paulo as invited guests, supported remotely by veteran member Peter Eisentraut. The duo compared the collaborative environment to a PostgreSQL "Commitfest," where technical papers are proposed, debated, and refined much like code patches.

Facetime at PGConf India

No less than 83 colleagues met at PGConf India in March! For some candid shots, as well as presentation assets for our sessions (including links to recordings), those can be found here: https://www.enterprisedb.com/event/pgconf-india-2026

PGConf Germany

Several colleagues presented sessions at PGConf DE. Peter Eisentraut talked about what the CRA (Cyber Resilience Act) means for PostgreSQL and its users, Charly Batista showed how to use machine learning to optimize PostgreSQL, and Bruce Momjian discussed what’s missing in Postgres (and why). Related, in his sponsor keynote Alvaro Herrera speculated about the next 10 years of the project. And Raphael Salguero Aragón shared strategies for migrating to PostgreSQL in the first place.

Gianni Ciolli presented a lightning talk as a follow-up to the session he gave at PGConf DE last year, and presented an update to his pgwtc extension, which he uses to analyze Bach’s music. Our Pinaz Raut, Charly Batista, Martin Marques, and Bruce Momjian also presented lightning talks.

For all slides: https://www.enterprisedb.com/event/pgconf-de-2026

Amsterdam colors Cloud Native

March was all about Kubernetes. Gabriele Fedi and John Long presented “Solving Operator Extensibility: A gRPC Plugin Framework for Kubernetes” at Cloud Native Rejekts, March 21, in Amsterdam, the Netherlands. In the days following several colleagues attended Maintainers Summit, and then the European KubeCon + CloudNativeCon.

Gabriele Bartolini was joined by Laurent Parodi from HSBC to talk about the bank’s journey, and adoption of the CloudNativePG project. Gabriele and Floor Drees presented “Beyond the DBaaS trap: Achieving Data Sovereignty With Kubernetes and CloudnativePG” at the Open Sovereign Cloud Day, a co-located event.

For slides, recordings, and pictures: https://www.enterprisedb.com/event/kubecon-cloudnativecon-eu

The 2026 Security Slam, an initiative aimed at “securing open source at the source”, culminated in an Awards Ceremony at KubeCon + CloudNativeCon. CloudNativePG finished in the top 3 achieving all 5 objectives. Pictured above are 5 of the project maintainers, and EDB engineers.

EDB’s PGConf.dev Preview

The EDB team is headed to Vancouver for PGConf.dev! Beyond the code, we’re focusing on the “Human Architecture” of the project, tackling contributor burnout, mentoring new developers through “Developer U,” and evolving the Community Code of Conduct.

Shoutout to our own Robert Haas and Jacob Champion for their leadership on the organizing committees! Will you be in Vancouver? Let’s grab a coffee in the hallway track.

Deployment of PostgreSQL Replica Cluster via Barman Cloud Plugin on CloudNativePG

This blog outlines the step-by-step process for setting up an EDB Postgres® AI for CloudNativePG replica cluster using the Barman Cloud Plugin for backups and WAL archiving. Or, part 1 of what Swapnil Suryawanshi, Principal Support Engineer, has in store for us: https://www.enterprisedb.com/blog/deployment-postgresql-replica-cluster-barman-cloud-plugin-cloudnativepg-part-1

The Subtle Shift: Understanding NOT VALID vs. NOT ENFORCED Constraints in PostgreSQL

PostgreSQL is famous for its data integrity, but as datasets grow into terabytes, the cost of maintaining that integrity can become a bottleneck. To solve this, PostgreSQL offers different “states” for constraints.

While many are familiar with NOT VALID, a new concept from the SQL:2023 standard has entered the conversation: NOT ENFORCED, which has been officially added to PostgreSQL 18. While they might sound similar, they serve very different roles in a database’s lifecycle. Amul Sul explains the difference on the blog: https://www.enterprisedb.com/blog/subtle-shift-understanding-not-valid-vs-not-enforced-constraints-postgresql-0

Validating the shape of your JSON data

On the blog Andrew Dunstan introduces a new PostgreSQL extension json_schema_validate that lets you validate JSON and JSONB data against JSON Schema specifications directly in the database.

The MySQL shadow

On his personal blog Bruce Momjian writes about how for much of Postgres’s history, it has lived in the shadow of other relational systems (and for a time even in the shadow of NoSQL systems).

Those shadows have since faded, but it is still helpful to reflect on that time.

Until next time

We hope you enjoyed this edition of the EDB Engineering Newsletter! Closing out with our Slack highlight of this month. Robert Haas will talk without interruption, but 74 weeks straight is a stretch, even for him...

Until next time,

The EDB Engineering Team

EDB Engineering Newsletter