Neues vom PostgreSQL Planet

On Tuesday, we had our last meetup of 2025, and we want to thank everyone who attended, and who supported us during 2025. An of course, very special thanks to Jay Miller, who present at both our first and last meetups! Everyone enjoyed the presentation, as well as pre-and post-conversations and pizza! Here as the presentation slides (Jay, thank you for sharing!):

The beautiful Old College building at the University of Edinburgh. Photo by LWYang from USA (CC BY 2.0).

I'm thrilled to announce that the PostgreSQL Edinburgh meetup is finally here! 🐘

Is your PostgreSQL database Feeling Sluggish? Are SQL statements taking more time than in earlier days? Are you experiencing performance cliffs (Unexpected, sudden drops in performance)? Are backups taking a long time to complete? Are you getting a bigger bill for storage? Are standby rebuilds and development refreshes becoming a herculean task? Are the auditors […]

Postgres 18, released on September 25, 2025, introduces an exciting set of performance improvements and new capabilities. Postgres has grown remarkably over the years, and with each major release has become a more robust, reliable, and responsive database for both mission critical and non-mission critical enterprise applications. I’ve been writing about these enhancements since the release, and today I’m diving into two more features that are especially useful in real-world applications.

When creating data models in PostgreSQL (or any other powerful relational database), several common problems can arise that hinder effective database design. One frequent issue is the incorrect normalization of tables, leading to redundant data, inefficient querying capabilities, or even simply plain wrong semantics. 

This post addresses one of the most common mistakes made by people new to the subject: Over-optimistic normalization. Especially developers relying too heavily on AI-generated data models are known to face this issue.

I recently watched a video about Starbucks and how their third place concept made them successful, and how their abandonment of it has caused problems. I started to realize that Postgres conferences also function as a "third place" — not home, not work, but somewhere people gather to see see old faces, new people, and to talk.

Everyone’s freaking out about AI taking their jobs. Meanwhile, I’m a DBA sitting in the corner thinking: “If programmers of the future are AI agents, companies will need 100x more human DBAs to clean up the mess in production.”

This rant blogpost is my attempt to explain why.

LLMs: Optimizing for the Next Token, Not for Reality

Let’s start with the core problem:

This year's final round of minor releases was released on schedule last week, with PostgreSQL 13 receiving its final update.

Do you need a PostgreSQL license for critical production use?
Short answer: No. 

The open-source PostgreSQL database is free to download, use, modify, and distribute. There are no per-CPU, per-core, per-socket, or per-instance license fees.

What a journey! I’m very happy to announce that all seven Google Summer of Code 2025 contributors successfully passed their final evaluations and made great contributions to the PostgreSQL community! 🎉

Back in May, I welcomed these talented people to our community. Now, six months later, I’m proud to celebrate not just the code they wrote, but the journey they made and the community members they’ve become.

The TLS server provides an alternative to using SSH for protocol connections to remote hosts.

In this demo setup, the repository host is named backup-srv, and the two PostgreSQL nodes participating in Streaming Replication are pg1-srv and pg2-srv. All nodes run on AlmaLinux 10.

If you’re familiar with Vagrant, here is a simple Vagrantfile that provisions three virtual machines using these names:

Just got home from KubeCon.

One of my big goals for the trip was to make some progress in a few areas of postgres and kubernetes – primarily around allowing more flexible use of the linux page cache and avoiding OOM kills with less hardware overprovisioning. When I look at Postgres on Kubernetes, I think there are idle resources (both memory and CPU) on the table with the current Postgres deployment models that generally use guaranteed QoS.

Berlin PostgreSQL Meetup on Thursday, November 13 2025 organised by Sergey Dudoladov and Andreas Scherbaum

Speaker: Josef Machytka

New York City PostgreSQL Meetup on Wednesday, November 12, 2025.

In this series, we talk about the many different ways you can speed up PostGIS. A common geospatial operation is to clip out a collection of smaller shapes that are contained within a larger shape. Today let's review the most efficient ways to query for things inside something else.

Frequently the smaller shapes are clipped where they cross the boundary, using the ST_Intersection function.

The naive SQL is a simple spatial join on ST_Intersects.

If you have ever generated a complex report in Odoo only to watch the loading spinner for minutes, you are not alone. One of our customers ran into exactly this scenario: their system ground to a near stall whenever they tried to compile business reports. After a systematic investigation, we achieved a 93 % performance improvement, but only by choosing the last resort: upgrading the instance’s resources.

This blog walks through the diagnosis, the dead ends we hit, and why scaling up was ultimately the right solution.

Learn how to use UUIDv7 today with stable releases of Python 3.14, Django 5.2 and PostgreSQL 18. A step by step guide showing how to generate UUIDv7 in Python, store them in Django models, use PostgreSQL native functions and build time ordered primary keys without writing SQL.

TL;DR - RegreSQL brings PostgreSQL's regression testing methodology to your application queries, catching both correctness bugs and performance regressions before production.

 You are trying to load a CSV file into a database, and almost inevitably, you run into an ERROR: value too long for type character varying 64 if you are using PostgreSQL.  Or SQL Error [1406] [22001]: Data truncation: Data too long for column 'Name' at row 1 if you are using MariaDB or MySQL.

Eight years ago, PostgreSQL introduced scram-sha-256 hashes for passwords.

Eleven months ago, MD5 was deprecated.

Yesterday, I got caught up with all this.

Some of this post will deal with how I fixed it, but mostly it is documenting (for myself) what I did. The fix covers several services and takes place over multiple days.

First, some background on why this change has come into focus for me.

Did you know a table can have no more than 1,600 columns? This blog article was inspired by a conversation Pierre Ducroquet and I had.

First, the documentation

The PostgreSQL documentation Appendix K states a table can have a maximum of 1,600 columns.