PostgreSQL Berlin February 2025 Meetup

Zalando is graciously hosting us for our first Meetup in 2025. And we are doing it a bit different: we have two rooms, and two talks in parallel. This gives you the opportunity to choose between two talks, for a total of four talks.

PLEASE BRING YOUR PHOTO ID
USE THE MAIN ENTRANCE TO BHW ON THE SIDE OF UBER ARENA

Agenda

Two sessions run in parallel in rooms codenamed "Auditorium" and "Festival"

18:30 - doors open: registration, networking, and drinks

19:00 - 19:05 Auditorium, "Event intro" by Sergey Dudoladov, Senior Database Engineer in Zalando.

Auditorium: "PostgreSQL performance in the AI Era" track

19:10 - 19:50 Bruce Momjian: Databases in the AI Trenches
19:50 - 20:10 Networking, Q&A, and pizza
20:10 - 20:25 Oleksandr Schulgin: SQL: The Good, The Bad and The Ugly, Lightning talk
20:30 - 21:00 Priyanka Chatterjee: Your Postgres DB performance Radar : pgBadger

Festival: "Running PostgreSQL at Scale: Lessons from the Trenches" track

19:15 - 19:45 Kaarel Moppel: Spot VMs and Postgres
19:45 - 20:10 Networking, Q&A, and pizza
20:10 - 20:25 Felix Kunde: "Doing PostgreSQL as a service so nobody cares about Postgres. Oh no ?" Lightning talk.
20:30 - 21:00 Alexey Kondratov: Serverless PostgreSQL: the journey from ~1s startup time to 10s and back

21:00 - 22:00 Networking, Q&A, and pizza
22:00 - The End

Abstracts

Databases in the AI Trenches

Artificial intelligence use has exploded, with much anticipation about its future. This talk explores many of the advances that has fueled this explosion, including multi-dimensional vectors, text embeddings, semantic/vector search, transformers, generative AI, and Retrieval-Augmented Generation (RAG). The talk includes semantic/vector search and rag examples. It finally covers how the valuable data stored in databases can be used to enhance AI usage.

Spot VMs and Postgres

This presentation explores common Postgres deployment options, introduces Spot VMs in general and sheds light on an avenue that only a few brave have dared to explore - running self-managed Postgres in the cloud using essentially ephemeral VMs.

SQL: The Good, The Bad and The Ugly, Lightning talk

In this talk, Oleksandr explores SQL as the primary application interface for database communication, reflecting on its role in modern development. He will delve into the good, the bad, and the ugly aspects of SQL, highlighting its strengths while addressing its limitations and challenges. With insights drawn from years of experience, Oleksandr will provide a balanced perspective on how SQL shapes database interactions and its impact on application design. This session aims to equip attendees with a deeper understanding of SQL’s potential and pitfalls.

Serverless PostgreSQL: the journey from ~1s startup time to 10s and back

In this talk we will fast-forward through the path from the first naive implementation of Neon Serverless Compute (Postgres), which struggled to scale, to the current state when we start hundreds of instances every minute with a p90 startup time under 1 second. We will discuss key optimizations along the way, including the pool of pre-provisioned VMs with fully automated pool size prediction and hit rate at startup above 99% in all regions.

About the speakers

Kaarel Moppel is a freelance PostgreSQL Consultant, with 20K+ hours of experience. Enjoys solving weird performance problems and hacking on tooling on spare time.

Note

We will again take notes who signs up and does not show up. This information will be used to provide seats to others at future events. Plain and simple: please only sign up if you plan to attend.