Sentry Lets AI Agents Build and Edit Dashboards
Sentry now offers agentic dashboard creation and editing, replacing Insights pages with cloneable templates and adding full CLI-based dashboard management.
Sentry has rebuilt its dashboard workflow around large language models, letting teams create and edit dashboards through the same agentic chat experience that powers its Seer AI debugger. The feature is in open beta and free for any organization with AI capabilities enabled, reducing dashboard creation from a manual, widget-by-widget process to a simple prompt.
The previously static Insights pages have been replaced with cloneable, editable 'Sentry-built Dashboards' covering common use cases like Web Vitals, Frontend Session Health, Backend Overview, and Mobile Vitals. While these templates can't be edited directly, they can be duplicated and then customized via an agent. Dashboards now also maintain a full revision history, so any AI- or human-made mistake can be instantly reverted.
For engineering workflows, Sentry added a complete dashboard command to its CLI, allowing dashboards to be listed, created, and modified straight from the terminal. With JSON output support, tools like Claude Code, GitHub Copilot, and Cursor can provision dashboards programmatically. Sentry demonstrated the approach with real cases—diagnosing a jest CI slowdown and testing a yabeda metrics integration—building targeted monitoring dashboards in minutes.
Overall, the update sharply lowers the cost of building observability views, letting teams spend less time assembling dashboards and more time acting on the insights they surface.