Microsoft Fabric Database Agent: Continuous Database Monitoring and Guided Investigation

The session introduces Microsoft Fabric Database Hub and Database Agent as an approach to database management that shifts from dashboard-and-alert monitoring to continuous observation and evidence-based issue reporting. It demonstrates how the agent detects meaningful SQL Server issues, provides root-cause context, and connects investigation to SSMS and GitHub Copilot workflows.

■ Database Agent concept and Fabric Database Hub
- Continuous observation surfaces issues with evidence instead of requiring constant dashboard checks
- Enabling the agent, selecting databases, and choosing scenarios such as blocking, long-running queries, indexes, CPU, memory, IO, and plan regression

■ Investigation and tooling
- Issue details include root cause, recommendations, and approval-gated actions such as killing a blocking session
- Opening SSMS with server, database, and query context, plus using GitHub Copilot agent mode and skills for deeper analysis

■ Estate-level workflows and roadmap
- GitHub Copilot chat can query database estate status and guide deeper diagnostics with approval steps
- Planned capabilities at Ignite, FabCon, and Build, with autonomy considered only after repeated verified success and auditable controls

It is suited for database administrators, developers, and data platform engineers evaluating AI-assisted database operations in Microsoft Fabric. Viewers gain an overview of how Database Agent supports detection, investigation, and guarded remediation, and can assess whether to try Fabric Database Hub and review the linked workload monitor.