EDB Postgres AI Overview - Release notes

Suggest edits

EDB Postgres® AI is a a new era for EDB. With EDB Postgres AI, customers can now leverage EDB's enterprise-grade Postgres offerings to support not just their mission critical transactional workloads, but also their analytical and AI applications. This also means that, in addition to the core transactional database releases you have come to expect from EDB, we will now be delivering regular updates to our analytics, AI, and platform capabilities.

This is the first quarterly release of EDB Postgres AI, delivering key functionality to support the platform's vision. This release includes analytical and vector database capabilities, single pane of glass management and observability for hybrid data estates, and an AI migration copilot.

EDB analytics and AI updates

Customers can now launch Postgres Lakehouse nodes using the EDB Postgres AI Cloud Service (formerly EDB BigAnimal) to get results of analytical (OLAP) queries much faster than from typical transactional Postgres. Postgres Lakehouse nodes are available now for customers using EDB Postgres AI - Hosted environments on AWS, and will be rolling out to additional cloud environments soon.

Postgres Lakehouse uses Apache DataFusion's vectorized SQL query engine to execute analytical queries 5-100x faster (30x on average) compared to native Postgres, while still falling back to native execution when necessary. Postgres Lakehouse nodes run either EDB Postgres Advanced Server or PGE as the Postgres engine, with data for analytics stored as columnar tables in object storage using the open source Delta Lake protocol. Customers can sync tables from transactional sources (initially, EDB Postgres AI Cloud Service databases) into Lakehouse tables in managed storage locations (initially, S3 object storage buckets).

Technical preview of EDB pgai extension

Customers can now access a technical preview of the new EDB pgai extension, which seamlessly integrates and manages AI data for enterprise workloads with EDB Postgres AI, to help understand your AI data directly out of the box. Built on top of Postgres vector data support, this tech preview enables Postgres to run LLMs and directly manage, process, search and retrieve AI data such as text documents or images to accelerate AI application development and operationalization across your company.

In this technical preview, you'll have the opportunity to explore the pgai extension and build AI-infused similarity search applications — for instance, a Retrieval-Augmented Generation (RAG) application using Postgres. RAG applications utilize a powerful combination of retrieval systems and language models to provide accurate and context-aware responses to user queries. Learn more and enroll in the tech preview here.

EDB platform updates

EDB Postgres AI Platform Agent release and platform support

As part of its initial release, the EDB Postgres AI agent enables users to connect self-managed Postgres deployments to the platform, to enable unified observability and management over hybrid data estates. Additionally, users will be provided with the Postgres database version and size (in MB) in the EDB Postgres AI Platform interface, with data collected from each database at a configurable level. Additionally, EDB Postgres All Platform is available on EDB-supported x86 Linux distros.

EDB Postgres AI database updates

EDB database server updates

As part of EDB's support for the open source community's quarterly release schedule, we completed PGE and EDB Postgres Advanced Server merge updates from the latest upstream PostgreSQL, including the following:

Database distributionsVersions supported
PostgreSQL16.3, 15.7, 14.12, 13.15 and 12.19
EDB Postgres Extended Server16.3.0, 15.7.0, 14.12, 13.15 and 12.19
EDB Postgres Advanced Server16.3.0, 15.7.0, 14.12.0, 13.15.21 and 12.19.24

EDB Postgres® Distributed 5.5 release enhancements

Read scalability enhancements

EDB Postgres Distributed users can now increase client application performance by spreading their read load across multiple nodes within a region. As a result, enterprises realize better support of read-heavy workloads by routing their read queries to a separate endpoint. The feature improves the former EDB Postgres Distributed process, where client applications could only use the lead node to route their application traffic via PGD Proxy, for both reads and writes, potentially impacting performance during peak times.


We now offer support for DETACH CONCURRENTLY commands for EDB Postgres Distributed (and all EDB database version types), which enables other SELECT queries to be executed on the parent table while the DETACH operation is underway.

For all the Q2 EDB announcements, visit the EDB blog.

Could this page be better? Report a problem or suggest an addition!