Hello Azure Community,

We’ve had a great month in March with some exciting updates to Azure Database for PostgreSQL  flexible server. From data replication with Microsoft Fabric to support for the latest PostgreSQL versions, these new features make it easier to build, manage, and scale your apps. Let’s take a quick look at what’s new.

Feature Highlights

1. Mirroring for Microsoft Fabric
2. AI Agents with Azure AI Agent Service
3. PostgreSQL Minor Versions
4. Version-less CMK
5. New Region: New Zealand North
6. Java SDK QuickStart
7. ADF Connector
8. Migration Enhancements
9. CLI & Portal Improvements

Mirroring for Microsoft Fabric - Public Preview

We’re excited to announce the public preview of Mirroring for Azure Database for PostgreSQL flexible server in Microsoft Fabric. This feature enables zero-ETL, near real-time replication from PostgreSQL flexible server to Fabric OneLake, making your operational data instantly available for analytics, AI, and reporting in Delta Lake format. Mirroring can be easily enabled in Azure Portal for your source databases, and a built-in azure_cdc extension to keep selected tables in sync with minimal impact on transactional workloads.

Explore how to get started here: Mirroring for PostgreSQL in Fabric

AI Agents with Azure AI Agent Service

Build an AI Agent today with Azure Database for PostgreSQL! We published a guide on how to use Azure Database for PostgreSQL and Azure AI Agent Service to develop intelligent AI agents. The guide walks through the significance of AI agents in automating complex workflows, how to set up your database as a tool for your agent and how to test the agent in Azure AI foundry.

For more details, check out this blog Build AI Agents with Azure Database for PostgreSQL and Azure AI Agent Service.

Latest PostgreSQL Minor Versions Supported

Azure PostgreSQL Flexible Server now supports the latest PostgreSQL minor versions:
17.4, 16.8, 15.12, 14.17, and 13.20.

These updates include critical security patches, performance improvements, and bug fixes. As always, updates are rolled out during scheduled maintenance windows to ensure your servers remain secure and up to date with minimal disruption.

Details: PostgreSQL Community Release Notes

Version-less Customer Managed Key (CMK) - Public Preview

Simplify management of auto-rotated keys in Key Vault. Using version-less keys removes the need to update the CMK information after an auto or manual rotation of the key in Key Vault.

For more details, check out the documentation - Versionless customer managed keys.

New Region: New Zealand North

We’re expanding again! Azure Database for PostgreSQL flexible server is now available in New Zealand North, giving customers in and around the region lower latency and data residency options. This continues our mission to bring Azure PostgreSQL closer to where you build and run your apps.

For the full list of supported regions, visit: Azure Regions  

QuickStart Guide for Java SDK

We’ve published a new Java SDK QuickStart guide to help developers quickly get started with Azure Database for PostgreSQL flexible server. The guide walks through basic setup, authentication, and server management using Java and Azure SDK libraries. Whether you're a backend developer or integrating PostgreSQL into a Java microservice, this guide gets you up and running fast.

For more details on using this SDK, please follow this link.

New Azure Data Factory (ADF) Connector

A new Azure Data Factory connector for is now generally available. This connector is designed with enhanced security features, including support for TLS 1.3, multiple SSL modes, and Service Principal (Entra) Authentication.

Whether you are migrating, transforming, or syncing data, this connector supports flexible data ingestion patterns and simplifies integration with other analytics or storage services in Azure. These improvements ensure added security when building data pipelines in Azure Synapse or Azure Data Factory when using Azure Database for PostgreSQL flexible server.

Learn more: Copy and transform data in Azure Database for PostgreSQL - Azure Data Factory & Azure Synapse | Microsoft Learn

Enhancements in Migration Service

The migration service now supports a broader set of PostgreSQL extensions, including pg_partman, postgis_topology, timescaleDB, and postgis_tiger_geocoder - ensuring smoother migrations for workloads using these popular extensions. It also supports migrating configuration tables defined by extensions, preserving critical metadata and settings during the migration process - similar to how pg_dump operates.

To give users more control, we've introduced configurable migration parameters directly in the server parameters section. These allow you to fine-tune migrations based on your data needs, reducing the need for manual intervention or support requests.

To learn more, refer to extensions migration

Improvements to CLI & Portal Experience

We’ve rolled out a set of important improvements to the Azure CLI and Portal to make managing PostgreSQL servers more intuitive, consistent, and scalable.

In CLI , we've added over 30 quality improvements to various command groups, including backup, firewall-rule, and more. New capabilities include support for index tuning, Fabric Mirroring, and the ability to specify the restore-time for the geo-restore command. Additionally, you can now list and scale Elastic Clusters. To learn more, read azure-cli updates.

In Portal, we've improved terminology, kept menus accessible regardless of server state, and enhanced existing functionality. You can now filter, group and sort your servers by properties like authentication method, vCores, storage size, restore point and more in the browse resources page.

Azure Postgres Learning Bytes 🎓

Maintaining hostname consistency with Virtual Endpoints during PITR

When recovering a server using Point-in-Time Restore (PITR), the default behavior assigns a new hostname - requiring application changes. Virtual Endpoints solve this by letting you assign a custom hostname that can be moved to the restored server, keeping your app connection string unchanged.

To use this:

• Create a virtual endpoint on your primary server and point your app to it.
• When you perform PITR, simply detach the endpoint from the old server and reattach it to the new one.
• Your application continues using the same hostname without needing any changes.

This simplifies disaster recovery, especially in production environments where DNS updates can be costly or error prone. To learn more, refer Using Virtual Endpoints for Consistent Hostname

Conclusion

That’s all for March! We hope these updates make your experience with Azure Database for PostgreSQL even better - whether you’re building AI-powered apps, moving data across services, or keeping your databases up to date and running smoothly. We’re always looking for ways to improve the service, and your feedback plays a big role in shaping what comes next. If you have suggestions, ideas, or questions, we’d love to hear from you.

📢 Share your thoughts here: aka.ms/pgfeedback

Thanks for being part of our growing Azure Postgres community. Stay tuned for more updates in April!