What Happened at PostgreSQL Conference Europe 2025

The fifteenth edition of PostgreSQL Conference Europe took place from October 22 to 24, 2025, in Riga, Latvia, bringing together database administrators, developers, and experts from the PostgreSQL community.

The conference has traditionally established itself as one of the leading events for open-source databases in Europe, and this year’s program offered rich content – from in-depth technical analyses to discussions about the community and project development.

PostgreSQL 18 – Main Focus of the Event

Many of the presentations were dedicated to the upcoming PostgreSQL 18 release and its improvements.

Key topics included:

• Asynchronous I/O (AIO) – introducing asynchronous input/output for reads, expected to improve performance under certain workloads;
• Improvements in constraint management – including new capabilities for PRIMARY KEY and UNIQUE constraints and better support for partitioned tables;
• Optimizations in locking mechanisms for environments with a high number of cores;
• Additional improvements in performance and stability.

AI, Vector Search, and Analytical Workloads

Reflecting current industry trends, the conference put strong emphasis on PostgreSQL’s role in AI environments.

Topics discussed included:

• Using PostgreSQL in the context of “agentic AI” and hybrid search;
• Optimizations when working with pgvector and managing large vector arrays;
• Comparisons with analytical solutions such as DuckDB and ClickHouse and discussion of when specialized tools are more appropriate;
• Operations, Kubernetes, and high availability;

For professionals managing production environments, solutions were presented for:

• Deploying PostgreSQL in Kubernetes environments;
• Managing high availability with Patroni;
• Disaster recovery using tools like pgBackRest and pg_basebackup.

Performance and Internal Architecture

Some sessions offered a deep dive into PostgreSQL’s internal workings, including:

• New mechanisms for vacuum and “eager scanning”;
• Latency analysis using wait events;
• Transparent Data Encryption (TDE).

Development and Data Modeling

For developers, best practices were presented for:

• Choosing between JSONB and the relational model;
• Query analysis using EXPLAIN and the new pg_stat_plans extension;
• Demonstrating the capabilities of Recursive CTEs, including an impressive example of implementing the Tetris game entirely with an SQL query;

Community and Project Sustainability

In addition to technical topics, the conference addressed issues related to compatibility standards, project history, and why PostgreSQL became the leading open-source database.

Special attention was given to the role of “invisible work” in the community – management, translations, event organization – as well as the topic of mental health among developers.

Among the most discussed presentations were sessions on AIO in PostgreSQL 18, analysis of PostgreSQL’s evolution compared to MySQL, and the demonstration of Tetris implemented entirely with SQL.

All conference lectures have been published online, providing professionals with the opportunity to explore in detail the latest developments and trends in the PostgreSQL ecosystem.