v0.4.0 Release Notes

Release date: February, 2026

Release version: v0.4.0

SynxDB Cloud v0.4.0 continues to enhance the platform’s capabilities in enterprise data management, AI integration, and operation experience. This release introduces AI agent applications, enhanced FoundationDB operations, and usability improvements across the platform, such as:

• AI agent applications

  • AI Bot - Data and AI Bot - Doc integration: Introduces SynxDB AI Bot - Data and SynxDB AI Bot - Doc, enterprise-class intelligent agent applications built natively on the SynxDB Cloud platform, enabling natural language data querying and document intelligence capabilities.

• Enhanced FoundationDB operations

  • FDB monitoring and visualization: Introduces comprehensive FoundationDB monitoring metrics and visualization capabilities in the DBaaS Admin Console, enabling proactive monitoring and troubleshooting of metadata operations.

  • FDB backup to object storage: Supports backing up FoundationDB data to object storage, providing reliable disaster recovery and data protection capabilities.

  • Multiple Coordinators with FDB: Enables creating multiple Coordinator nodes when using FoundationDB as metadata storage, improving availability and scalability.

  • TpServer with FDB: Supports creating TpServer nodes when using FoundationDB as metadata storage, extending HTAP capabilities to FDB-based deployments.

• Resource management enhancements

  • Fuzzy search: Adds fuzzy search capabilities across various resource management interfaces, improving usability and search efficiency.

  • Database Object management: Provides unified management of databases, tablespaces, and related database objects through the DBaaS Admin Console, streamlining administrative workflows and improving operational efficiency.

  • Resource usage limits: Introduces resource usage limits for better resource allocation and cost control in multi-tenant environments.

New features

AI agent applications

SynxDB Cloud v0.4.0 introduces SynxDB AI Bot, an enterprise-class suite of intelligent agent applications natively built on the SynxDB Cloud platform. This suite integrates cutting-edge Agentic AI technology with scalable distributed data processing capabilities, delivering end-to-end solutions for data analytics and document intelligence.

AI Bot - Data integration

SynxDB AI Bot - Data empowers business teams to interact with databases using natural language, enabling instant data querying and analysis without requiring SQL expertise. Key capabilities include:

• Natural language querying: Enables users to query databases using plain English or Chinese, maintaining conversation context for seamless analysis.

• Smart SQL generation: Automatically converts natural language into optimized SQL queries, supporting complex multi-table joins and aggregations.

• Interactive visualization: Instantly generates dynamic charts (Line, Bar, Pie, Scatter, Heatmaps) based on query results for intuitive data exploration.

• Automated reporting: Compiles comprehensive analysis reports with executive summaries and visualizations, exportable to PDF, HTML, or Markdown.

• Attribution analysis: Diagnoses root causes by analyzing factors contributing to metric changes, such as identifying reasons for sales fluctuations.

For more details, see SynxDB AI Bot - Data.

AI Bot - Doc integration

SynxDB AI Bot - Doc provides enterprise knowledge management capabilities for unstructured documents and multimedia content. Key features include:

• Multi-modal knowledge base: Ingests and manages diverse unstructured data formats including PDFs, Word, Excel, images, and audio/video.

• RAG-driven Q&A: Utilizes Retrieval-Augmented Generation to deliver precise, context-aware answers with citations from enterprise documents.

• Visual data workshop: Provides a low-code, drag-and-drop workflow builder for custom document parsing, splitting, and vectorization pipelines.

• Semantic & hybrid search: Combines vector and full-text search to retrieve relevant documents and multimedia assets based on user intent rather than just keywords.

• High-fidelity parsing: Extracts text, tables, and images from complex documents while preserving original layouts for accurate knowledge processing.

For more details, see SynxDB AI Bot - Doc.

Enhanced FoundationDB operations

This release significantly enhances FoundationDB operations, providing better observability, reliability, and operational flexibility for deployments using FoundationDB as metadata storage.

FDB monitoring and visualization

Introduces comprehensive FoundationDB monitoring metrics and visualization capabilities directly within the DBaaS Admin Console. Administrators can now monitor key FoundationDB metrics including performance indicators, health status, and operational statistics. This enhanced observability enables proactive troubleshooting and ensures the stability of metadata operations in FoundationDB-based deployments.

For more details, see View FoundationDB details.

FDB backup to object storage

Supports backing up FoundationDB data to object storage, providing reliable disaster recovery and data protection capabilities. This feature enables administrators to create and manage FoundationDB backups, ensuring data durability and enabling point-in-time recovery scenarios. By leveraging object storage for backups, organizations can achieve cost-effective and scalable backup strategies.

For more details, see View FoundationDB details.

Multiple Coordinators with FDB

Enables creating multiple Coordinator nodes when using FoundationDB as metadata storage. This capability improves system availability and scalability by distributing coordinator workloads across multiple nodes. Multiple coordinators provide redundancy and enable better handling of concurrent administrative operations.

TpServer with FDB

Supports creating TpServer nodes when using FoundationDB as metadata storage. This extends HTAP (Hybrid Transactional/Analytical Processing) capabilities to FDB-based deployments, enabling high-concurrency transactional processing and streaming data ingestion in FoundationDB environments.

Resource management enhancements

Fuzzy search

Adds fuzzy search capabilities across various resource management interfaces in the DBaaS Admin Console, including:

• Alert rules search by name

• Bucket search by name

• Contact groups search by group name

• Environment specifications search by ID and name

• Configuration parameters search

• Region name search

• Operation users search by user name

• LDAP configuration search by name

Fuzzy search improves usability by allowing users to find resources even with partial or approximate matches, reducing the need for exact spelling and improving search efficiency.

Database object management

Provides comprehensive database object management capabilities through the DBaaS Admin Console, enabling administrators to efficiently manage databases, tablespaces, and related objects within accounts.

Database management: Supports creating, viewing, and deleting databases within accounts. Administrators can view database details including OID, name, owner, and perform deletion operations. This feature streamlines database lifecycle management and provides centralized visibility into database resources.

Tablespace management: Enables creating and managing tablespaces to optimize storage distribution. Key capabilities include:

• Tablespace creation: Create tablespaces with configurable names, owners, and storage buckets.

• Privilege management: Grant and manage user access to specific tablespaces, enabling fine-grained access control.

• Tablespace deletion: Safely remove tablespaces when no longer needed.

For more details, see Manage Databases and Manage Tablespaces.

Resource usage limits

Introduces resource usage limits for better resource allocation and cost control in multi-tenant environments. Administrators can configure limits on resource consumption to ensure fair resource distribution and prevent resource exhaustion. This feature supports policy-based resource management and enables more predictable performance characteristics.

Storage engine improvements

Cloud storage engine optimizations

Enhances Cloud storage engine performance with several optimizations:

• IO optimization for discrete columns: Optimizes IO reads for multiple discrete columns using io_uring, significantly improving query performance for columnar data access patterns.

• Prefetch optimization: Implements prefetching of the next group when scanning Cloud files, reducing I/O wait times and improving sequential scan performance.

• Docker image enhancement: Adds liburing library to the Docker image to support Cloud io_uring optimizations.

Query and transaction optimizations

Improves query performance and transaction handling:

• Lock optimization in COPY: Reduces lock operations in COPY commands using hash tables, improving concurrent COPY performance.

• Cloud manifest management: Enhances manifest deletion on relation drop and refactors snapshot handling for improved efficiency and reliability.

• Cloud transaction handling: Records invalid files at transaction abort, improving transaction rollback reliability and data consistency.

• Shared scan deduplication: Removes redundant Shared Scan operations after deduplication subqueries, optimizing query execution plans.

• Hashtable memory management: Automatically re-creates hashtables when memory usage exceeds limitations, preventing memory-related performance degradation.

Bug fixes

• Fix ANALYZE partitioned table coredump issue.

• Skip cdb_setup/cdb_cleanup for auxiliary background workers.

• Fix frozen insert catalog table for FoundationDB.

• Fix regress_cloud regression and unstable tests.

• Set omit_location correctly for consistent error behavior.

• Fix external table check and report errors.

• Fix gawk regexp escape sequence warning.

• Fix cdbgang_createGang_async timeout retry logic.

• Fix PAX compile issues and LLVM errors.

• Rename incremental nonterminal to opt_incremental in parser.

• Check create warehouse/tpserver privilege correctly.

• Fix TDE encrypt buffer context not using shared memory.

• Fix OuterParam as UDF args in subquery tlist.

• Fix UtilityContextClean to cleanup relation_oids.

• Fix missing WAL logging during relation data block extend.

• Remove PG-14 rebase artifact and fix dblink_get_conn in dblink extension.

• Fix incorrect column number in funcTupleDesc.

• Add pg_buffercache_usage_counts() function to pg_buffercache extension.

• Fix assertion failure when running psql \dit+.

• Prevent modification of profiles that are already bound to resources.

• Fix insufficient disk directory permissions that prevented FoundationDB service from starting.

• Encrypt pseudo console user passwords using salt.

• Encrypt storage of object storage secret keys.

• Fix various English spelling errors on pages.

• Fix loading screen not supporting i18n.

• Notify users that resource names are automatically converted to lowercase.

• Improve cloud manifest management by deleting manifests on relation drop and refactoring snapshot handling.

• Improve cloud transaction handling by recording invalid files at transaction abort.

• Optimize PAX IO reads for multiple discrete columns using io_uring.

• Optimize PAX file scanning with prefetch of next group.

• Reduce lock operations in COPY commands using hash tables.

• Improve hashtable memory management by re-creating hashtables when memory exceeds limitations.

• Remove redundant Shared Scan operations after deduplication subqueries.

• Port UDP motion layer fixes to udp2.

• Modify recycle bin catalog tables with recycleid field.

• Change ufile unlink log level for better logging.

• Update cyclonedx-maven-plugin version to 2.9.1.