Database Agent
An intelligent monitoring component built on top of Database Hub. It reads database telemetry to produce explainable diagnostic evidence and proactively identifies issues such as blocking and performance degradation.
Timeline (last 90 days)
The database Agent is an intelligent monitoring component built on top of Database Hub that forms explainable diagnostic evidence by reading database telemetry data (such as Query Store, DMVs, etc.), rather than relying on preset threshold alerts.
Enabling the database Agent requires a tenant administrator to search for Database Hub in the Microsoft tenant admin portal and turn on two separate switches, Hub and Agent; both are required.
After enabling the database Agent, there is a roughly 15-minute delay before it takes effect.
The current version of the database Agent supports only SQL databases, with Postgres and Cosmos being rapidly onboarded.
The database Agent can automatically discover issues such as blocking, query performance degradation, index tuning needs, and resource pressure, and proactively report them to administrators.
The current version of the database Agent is only responsible for discovering and detecting issues; investigation and remediation still require SSMS and Copilot.
In the demo, Patrick LeBlanc artificially constructed a blocking scenario using WAITFOR DELAY, and the Agent accurately captured the problem and directly provided the lead blocker query causing the blocking.
Clicking the issue details link automatically launches SSMS with the current server and database context and authenticates via Microsoft Entra ID, with no need to manually enter connection information.
Patrick LeBlanc said the 'north star' for the database Agent product is to give the Agent memory, and that the Day one goal is not full autonomy but gradually earning user trust through credible evidence.
Patrick LeBlanc revealed that there will be relevant announcements about the database Agent at the upcoming Ignite conference, but he could not disclose details for now.
All Facts (10)
The database Agent is an intelligent monitoring component built on top of Database Hub that forms explainable diagnostic evidence by reading database telemetry data (such as Query Store, DMVs, etc.), rather than relying on preset threshold alerts.
50%UnverifiedEnabling the database Agent requires a tenant administrator to search for Database Hub in the Microsoft tenant admin portal and turn on two separate switches, Hub and Agent; both are required.
50%UnverifiedAfter enabling the database Agent, there is a roughly 15-minute delay before it takes effect.
50%UnverifiedThe current version of the database Agent supports only SQL databases, with Postgres and Cosmos being rapidly onboarded.
50%UnverifiedThe database Agent can automatically discover issues such as blocking, query performance degradation, index tuning needs, and resource pressure, and proactively report them to administrators.
50%UnverifiedThe current version of the database Agent is only responsible for discovering and detecting issues; investigation and remediation still require SSMS and Copilot.
50%UnverifiedClicking the issue details link automatically launches SSMS with the current server and database context and authenticates via Microsoft Entra ID, with no need to manually enter connection information.
50%UnverifiedPatrick LeBlanc said the 'north star' for the database Agent product is to give the Agent memory, and that the Day one goal is not full autonomy but gradually earning user trust through credible evidence.
50%UnverifiedIn the demo, Patrick LeBlanc artificially constructed a blocking scenario using WAITFOR DELAY, and the Agent accurately captured the problem and directly provided the lead blocker query causing the blocking.
50%UnverifiedPatrick LeBlanc revealed that there will be relevant announcements about the database Agent at the upcoming Ignite conference, but he could not disclose details for now.
50%