Sentry

Connect Sentry and your production errors become a source of work. Errors come in, pass a noise gate, and the ones worth fixing turn into tickets — Leroy reads the stack trace, finds the cause, and opens a pull request. Defaults are conservative, so you get fixes for real signals, not a flood of tickets for every blip.

01 What it does

Sentry catches errors in production. Leroy turns the ones that matter into fixes. The flow:

• Errors arrive from your connected Sentry projects.
• A noise gate filters them. One-off blips, low-frequency noise, and non-actionable events are held back. Only issues that clear the gate move forward.
• The ones worth fixing become tickets. A promoted error becomes a ticket in Leroy with its stack trace and context attached.
• Leroy investigates and ships a PR. It reads the stack trace, traces it to the cause in your code, writes the fix, and opens a pull request.

02 Connect it

You connect Sentry from the Integrations screen in the Leroy app. Exact steps live there, but the shape is:

1. Open Integrations → Sentry in the Leroy app and start the connect flow.
2. Authorize Leroyfor your Sentry organization through Sentry's install prompt.
3. Choose the projects whose errors should flow into Leroy.
4. Confirmand you'll return to Leroy with Sentry connected. New errors start flowing through the noise gate from here on.

03 What it unlocks

With Sentry connected:

• Production errors fix themselves into PRs. A real, recurring error can become a reviewable pull request without anyone filing a ticket by hand.
• Stack-trace-aware work. Leroy starts from the trace and the failing context, so the fix targets the actual cause.
• Oversight you configure. Your policy gates can treat Sentry-sourced work differently — for example, asking before it picks up or ships error fixes — independent of your other sources.