AI Assistant for Database Monitoring
Use the AI Assistant to investigate database health and performance.
The AI Assistant correlates database server health, query performance, wait states, blocking sessions, and execution plans to help you investigate database performance issues.
Prerequisites
Before you begin, make sure that:
-
The AI Assistant is available in your realm, active for your organization, and accessible to you.
-
Your organization has an active AI Assistant entitlement.
-
Your database instances send telemetry through a supported receiver: SQL Server, Oracle Database, PostgreSQL, or MongoDB.
-
Your user role has permission to view Database Monitoring and use the AI Assistant.
Limitations
-
While an in-context investigation is processing, other in-context investigation buttons are temporarily unavailable.
-
The AI Assistant analyzes active and recent telemetry within your organization's configured data-retention window.
Investigate database health from an Average Active Sessions chart
Start an investigation from an Average Active Sessions (AAS) chart to examine session distributions and identify wait classes or queries that contribute to a database health issue. The AI Assistant correlates database server resource metrics, such as CPU, memory, and disk I/O, with query performance data and wait states.
Select Explain with AI from the chart to open an investigation with relevant context.
Investigate a time range
Select a time range on a metric chart to focus an investigation on a database performance issue during that period. The AI Assistant uses the selected time range and related metric dimensions as context.
Select Explain with AI to open the investigation.
Investigate a top query
Start an investigation from a top query table to examine a slow-running SQL query, its performance history, and related wait events.
Select Explain with AI for the query you want to investigate.
Investigate from a query flyout
Start an investigation from a query flyout to provide context about the selected database instance, time range, query signature, and metric dimensions.
Select Explain with AI to open an investigation with the query context.
Analyze an explain plan
Start an investigation from an explain plan to analyze its operations and identify potential query bottlenecks, such as costly scans or inefficient joins. Explain plan analysis works for all supported databases.
Select Explain with AI from the explain plan.
Ask follow-up questions
Continue an investigation with a suggested follow-up prompt or enter your own question. For example, ask which session blocks a query, which wait event contributes the most latency, or whether an index might resolve a full table scan. The AI Assistant can suggest remediation steps, such as adding a missing index.