PGconf.de 2024 in Munich

As the two talks I had submitted did not get picked for this year´s PGconf.de 2024 in Munich, my prep for the conference was all about booking train tickets from Düsseldorf to Munich. Being an early bird definitely paid off - Bernd Helmle and I enjoyed first class luxury, while our procrastinating colleagues ended up paying more for second-class seats.

The Thursday journey was pretty smooth, with the conductor proudly announcing our punctual arrival - surprised it didn't make it to the news with Deutsche Bahn's track record. I even managed to get a good amount of work done, working on CYBERTEC's PostgreSQL Enterprise Edition Docker images that we want to publish soon.

The day before the event

Some colleagues arrived throughout the afternoon and we went for a "relaxing" 7 km walk to the city center before heading to the Wirtshaus Eder where we had our company dinner. It turned out to be the same restaurant the conference organizers had chosen for the speakers' dinner, which made it simple for our colleagues giving talks—they just had to decide whether to sit to the left of the right side of the main entrance. The food and service were great, and of course, nothing compares to Bavarian beer.

Conference day on Friday, 12th of April

The opening session on Friday started with a tribute to Simon Riggs, a cherished visitor and speaker at many past PGconf.de conferences, who had tragically passed away recently in a plane accident..

The first talk I attended was by Michael Banck, comparing the three major PostgreSQL cloud hosters - Amazon RDS, Google Cloud SQL and Microsoft Azure Postgres. An important highlight from the talk was that all three offer a (mostly) solid service, with Amazon being the most advanced one, and Azure having made significant progress over the past three years. This is also visible in the increasing number of community members working at these companies contributing to PostgreSQL.

I skipped the next session, chatting with Peter Eisentraut about community matters and PostgreSQL extension ecosystems.

Later, Laurenz Albe discussed security attacks on PostgreSQL. I was familiar with most of the attack vectors, but it was helpful to see a summary about real-world issues.

After lunch, there wasn't a session that particularly caught my interest based on the abstracts. I picked Wolf Beckmann's talk about practical usage of ChatGPT and databases. I was pleasantly surprised—the talk was very informative and the speaker's enthusiasm really came through and inspired the audience. I definitely have to give the API of these AI services a try and I can even see some potential uses for CYBERTEC. This talk ended up being the highlight of the conference for me, and I had a great chat with Wolf in the hallway that extended well into the next session.

In the next slot, Julian Markwort presented on 15 pitfalls and solutions for PostgreSQL replication. Some were surprisingly insightful, especially after having seen a lot of support cases over the years.

The conference wrapped up with Ants Aasma discussing aspects that PostgreSQL's "explain" doesn't reveal. After highlighting three missing performance metrics, he also proposed a list of eight more items that could be beneficial for performance debugging. This talk could provide useful ideas for PostgreSQL 18 development and beyond.

Several friends and colleagues stayed on until the next day, so we went for dinner in the Augustiner Bräustuben with excellent food and beer once again. Unfortunately, the train back on Saturday encountered some delay on the way to Cologne and Düsseldorf so we missed the connection there, but made it home eventually.

Overall, really enjoyed the conference, and believe it would be even better if expanded to two days. This would provide opportunity for more meaningful interactions and discussions between sessions.