AI-Detected Issues

Learn about AI-detected issues and how Sentry uses AI to find problems in your traces.

Sentry analyzes your traces and logs using AI to detect issues that traditional pattern-matching detectors can't catch. Instead of relying on hardcoded rules, AI Issue Detection identifies problems across broader trace context, using a trained model to create issues that closely match designated categories.

AI Issue Detection classifies detected issues into the following categories:

Issue CategoryDescriptionPossible Issue Title
HTTPIssues related to HTTP requests and responsesInefficient HTTP Requests, Degraded HTTP Operation, Failed HTTP Operation
DatabaseIssues related to database queries and operationsInefficient Database Queries, Degraded Database Operation
Runtime PerformanceGeneral runtime performance problemsBlocking Operation, Degraded UI Performance
SecurityPotential security concerns detected at runtimePotential Security Leak, Potential Security Risk
Code HealthMisconfigurations or usage of deprecated featuresConfiguration Warning, Deprecation Warning

AI Issue Detection is available to organizations with the Early Access feature enabled on a paid account. Organizations on a Business plan will have more of their traces analyzed for AI-detected issues.

You must also have the "Show Generative AI Features" toggle enabled in Organization Settings > General.

Under Project Settings > Performance, the AI Issue Detection toggle controls whether Sentry runs AI issue detection at all. If toggled on, you can enable or disable each category individually.

All toggles are enabled by default.

When viewing an AI-detected issue, the span evidence section shows:

  • Transaction — The transaction where the issue was detected (links to the transaction summary).
  • Explanation — A description of what the problem is.
  • Impact — The potential impact of the issue.
  • Evidence — Supporting detail from the trace.
  • Trace — A preview of the trace, and a link to the full trace view.
Was this helpful?
Help improve this content
Our documentation is open source and available on GitHub. Your contributions are welcome, whether fixing a typo (drat!) or suggesting an update ("yeah, this would be better").