What's new in IBM Db2 Genius Hub 1.1.3.0
Platform and administration enhancements
This release includes enhancements to platform support, authentication, query management, alert management, and connection organization.
- macOS support
Added support for macOS platform.
- Kerberos authentication support
IBM® Db2® Genius Hub now supports Kerberos authentication for secure single sign-on (SSO) in enterprise environments. Administrators can configure Kerberos realms, Key Distribution Centers (KDC), service principals, and map LDAP groups to roles. The configuration wizard provides step-by-step guidance with built-in testing capabilities. Supports both Active Directory and OpenLDAP environments.
- Registry guideline recommendations for query tuning
Query tuning now provides embedded guideline recommendations by using Db2 registry variables to optimize Columnar table performance (database versions 12.1.3 onwards). These guidelines can be appended directly to queries without requiring global configuration changes, significantly improving execution time.
- SQL workbench export privilege control
Console Admins can now control the Export to csv/xlsx capability in SQL workbench through the Roles and Privileges feature. The SQL workbench Export privilege is enabled by default for all roles to make sure backward compatibility, and can be disabled for Database Admin or Database User roles to prevent exporting query results to the local file system.
- Bulk alert deletion in notification center
Delete all alerts that match the selected filter criteria in the Notification Center, simplifying alert management and reducing manual effort.
- Filtering of key insights
Filter key insights by selecting a database tag, such as PureScale. The connections associated with the selected tag are displayed.
- Grouping connections
Group connections by Status or Alerts to organize them under expandable categories, such as Available, Disconnected, Critical, and Warning, with the number of connections displayed for each group.
Agentic AI features
This release introduces new agentic AI capabilities that enhance database diagnostics, automation, monitoring, and operational assistance across IBM Db2 environments.
1. Enhanced DBA expert agent
Dynamic Context agentic Loop - AI-powered senior database administrator that autonomously performs tasks, investigates, and resolves Db2 operational issues by using SMAP documentation.
Key Capabilities:
- Evidence-driven diagnostics - Forms hypotheses, validates with live telemetry, and iterates until root cause is identified
- SMAP-powered intelligence - Searches IBM Docs and Smart Agent Protocols (SMAP) documentation to understand error codes, symptoms, and proven resolution steps
- Human-in-the-loop safety - Requires explicit approval for critical operations (SQL execution, configuration changes, DDL operations)
- Multi-tool orchestration - Coordinates docs search (SMAP docs), telemetry tools, Safe SQL, and remote OS diagnostics
2. Run safe SQL
Security-hardened SQL execution layer enforcing read-only operations through multi-layered
validation by using sqlglot parsing.
Protection Features:
- DML/DDL/DCL/TCL blocking - Prevents
INSERT,UPDATE,DELETE,CREATE,ALTER,DROP,GRANT,REVOKE, and transaction control statements - UDF protection - Validates all function calls against an allowlist of safe Db2 built-in functions to prevent malicious user-defined functions
- Locking clause detection - Blocks
FOR UPDATEclauses that can acquire locks; allowsFOR READ ONLYorFOR FETCH ONLYonly - SQL injection prevention - Uses
sqlglotparser to properly handle comments, string literals, and complex SQL syntax - Row limit enforcement - Caps results at 25 rows (MAX_SAFE_SQL_ROWS) to prevent memory overflow and context window exhaustion
- Query tagging - Prepends
/* safe_db2 */comment for identification in Db2 monitoring tools
3. RemoteOS API tool
OS-level diagnostics accessing logs and cluster health beyond SQL telemetry.
Capabilities:
-
Db2 diagnostic logs (db2diag.log) -
Retrieves diagnostic log entries from remote OS host with time window filtering, severity levels
(
INFO,WARNING,ERROR,CRITICAL), and natural language time expressions ("2 hours ago", "yesterday", "last week") - Cluster health monitoring - Retrieves cluster type information, instance status, domain details, and quorum health for comprehensive cluster visibility
4. Controlled write actions
Human-in-the-Loop (HITL) approval mechanism for critical database modifications.
Three Control Areas:
-
Update database configuration (update db cfg)
- 16 Allowlisted parameters - LOCKTIMEOUT, LOCKLIST, MAXLOCKS, CUR_COMMIT, AUTO_RUNSTATS, WLM_AGENT_LOAD_TRGT, LOGSECOND, SORTHEAP, SHEAPTHRES_SHR, and more
- Pre-execution validation - Enforces parameter ranges, validates
AUTOMATIC/ON/OFFvalues, blocks CAS deployments, shows current versus deferred values - Restart warnings - Automatically detects parameters that require database restart (LOCKTIMEOUT, CUR_COMMIT) and warns users before approval
-
Force application (force application tool)
- Connection termination - Forcibly disconnects applications by handle with immediate rollback of uncommitted work
- Application validation - Verifies that application exists through
MON_GET_CONNECTIONSbefore attempting off force, prevents invalid handle errors - Rich context display - Shows application name, auth ID, connection start time, and active duration in approval message
-
Table reorganization
- Two operation types -
INPLACE CLEANUP OVERFLOWS(fast online fix for F1 violations) andRECLAIM EXTENTS(columnar table space reclamation) - Identifier sanitization - Regex validation blocks SQL injection through schema/table names, only allows alphanumeric and underscores
- Impact warnings - Notifies users of potential performance impact during reorganization,
recommends
ALTER TABLESPACE REDUCE MAXafter completion
- Two operation types -
5. Token optimization
Multi-layered approach to reduce LLM costs and improve response times.
6. Db2 Genius MCP
MCP Server providing AI assistants (Claude Code, Bob, Cursor, or customers' native AI agents) secure, tool-based access to Db2 monitoring, diagnostics, tuning, and historical performance analysis.
Key Tools:
- Query tuner
- Execute safe SQL
- Explain plan analysis
- WLM queuing analysis
- Db2 vector store semantic search
- Historical data analysis
- Change analysis
- Around 30 tools
7. New SMAPs
Refreshed private knowledge base with 25+ new SMAP documents covering enhanced troubleshooting scenarios and operational procedures.
8. PureScale cluster health checks
Comprehensive cluster health diagnostics via Remote OS API and SMAP-guided validation.
Commands:
- db2instance -list - Lists all Db2 instances on the PureScale cluster with status and version information
- db2cm -list -alert - Retrieves active Cluster Manager alerts for proactive issue detection and high-availability monitoring
- db2cm -list -domain - Shows cluster domain configuration including member nodes, CF (Cluster Caching Facility) servers, and network topology
- db2cm -list -quorum - Displays quorum device status and voting configuration for split-brain prevention
SMAP-Guided health checks:
- Automated validation - Runs all four PureScale commands and validates output against SMAP documentation patterns
- Anomaly detection - Identifies cluster state inconsistencies, failed members, CF connectivity issues, and quorum violations
- Python3 implementation - Native Python health check program by using subprocess execution with timeout protection and error handling
9. Additional Enhancements
- 25+ New SMAP Documents
- Expanded knowledge base for improved troubleshooting and operational guidance.
- Agent Harness and Improved Accuracy
- Enhanced agent accuracy and support for various deployment types.
- New Short- and Long-term Memory Tools
- Implemented user preferences and long-term recall capabilities.
- Improved Knowledge Search Accuracy
- Enhanced conversational search by retrieving documents more specific to instance types, deployment types, and version.
- Space Management Range Partition Tables Support
- Extended support for reorg checks and space reclamation opportunities for range partition tables.
- Setup HADR SMAP and TODO List Based Tasks
- SMAP document for setting up HADR leveraged by expert agent with a Todo list-based approach to solve the task.