PostgreSQL Usergroup NL Meetup - The Winter Edition

Dear all,

With winter approaching, we are happy to invite you to the next edition of the PostgreSQL User Group NL Meetup on December 3rd 2026 in Amsterdam.

This time we have again been kindly invited by JetBrains to host the event at their Amsterdam office. We are very appreciative of this gesture and would like to thank them for supporting the PostgreSQL community with the venue and the hospitality.

On the agenda we have two excellent speakers. First up is Jelte Fennema-Nio, PgBouncer maintainer and long-time Postgres contributor, who will share the interesting, funny and often horrifying lessons he’s learned building “Postgres compatible” servers. Next on the program is Nils Dijk, who will take us through the design of Azure HorizonDB and what it takes to disaggregate storage from compute for PostgreSQL.

As usual, we have reserved some space in the program for anyone who wants to do a lightning talk. If you are interested, please contact us via a personal message on meetup.com.

We are very grateful for the contributions of both presenters as well as our host, to make this all possible for the PostgreSQL community in the Netherlands.

We hope to meet many of you at the meetup — a moment where PostgreSQL users can connect, share knowledge, and discuss the latest developments in the PostgreSQL ecosystem.

We look forward to a great meetup!

Feike Steenbergen
Gerard Zuidweg

***

Meetup Program

17:30 – 18:25 — Welcome, Beer & Bites

18:30 — Opening

18:30 – 19:10 — Just enough Postgres compatibility to keep your sanity – Jelte Fennema-Nio

Abstract:
Postgres has become the most popular database — which is why many other databases claim to be “Postgres compatible”. But the hard truth is: only PostgreSQL is fully “Postgres compatible”. Everyone else is just a clever impostor attempting to fool most users. But given enough users, there’s always one who finds a new compatibility bug.
Jelte has spent years perfecting this art of deception, having implemented the Postgres protocol twice: first as a maintainer of the PgBouncer connection pooler, and then while building the “Postgres Endpoint” for MotherDuck (a data warehouse based on DuckDB). In this talk he’ll share the interesting, funny and often horrifying lessons he’s learned about building Postgres-compatible servers.
What you’ll learn in this talk:

• How to build your own “Postgres compatible” database with just a few lines of Go code
• A deep understanding of the Postgres protocol
• Why the line between the protocol and the SQL dialect is kinda blurry
• How clients and ORMs (mis)use the system catalogs in pg_catalog and information_schema
• When to say: “This is enough!”

About Jelte:
Jelte Fennema-Nio is a PgBouncer maintainer and Postgres major contributor. He currently works at MotherDuck, a data warehouse company built on DuckDB, where he combines Postgres and DuckDB in various ways. Before that, he worked for 5 years at Microsoft on Citus, Postgres and PgBouncer.

— (Very) Short Break —

19:15 – 19:55 — Disaggregating PostgreSQL Storage from Compute with Azure HorizonDB – Nils Dijk

Abstract:
Building a completely new storage layer for a transactional database is not something many people will advise you to do — quite the contrary. However, when the upsides are big, it is well worth the effort. Having the benefit of a mature environment like Azure, and a team with decades of experience in database internals, makes a seemingly impossible task doable. In this talk we will walk through the design of HorizonDB, compare it to the usual suspects, and highlight the different trade-offs we have made. We will touch on Aurora, Socrates, Neon and many Postgres internals, but no prior experience is required. As a bonus, we will discuss how to secure untrusted C code in a multi-tenant enterprise environment without reimplementing Postgres’ redo subsystem.

About Nils: Nils Dijk is a Principal Software Engineer currently enjoying a multi-retirement. Before stepping away, he was part of the team behind Azure HorizonDB, disaggregating storage from compute for PostgreSQL. Although today he works solely on Postgres-related projects, he started his career in the mid-2010s and was, as a side effect, clouded by the NoSQL philosophy — until that came to an abrupt end while working on the dispatch systems at Uber. Even though the system of record there was NoSQL-based, the actual business value came from the SQL layer built on top. With this experience, he set out to scale SQL databases without giving up their benefits to the business, which led him to join Citus Data, where he worked on everything from the query planner to cluster management — fulfilling the vision of a seamlessly scaling SQL database.

19:55 – 20:00 — Lightning Talks

20:00 – 20:30 — Wrap-up and drinks