Automation 12/30+: Put an AI agent inside Sentry

Description and use

An AI agent reads Sentry alerts, writes a root‑cause sketch, creates or updates a GitHub issue, and stores the prompt and transcript as a CI artifact.

Problem

Sentry alerts generate repetitive manual triage: developers open stack traces, guess causes, and close tabs, repeating the process daily.

Outcome

Automated root‑cause sketches reduce manual triage, prevent duplicate issues, and provide traceable reasoning via CI artifacts.

Workflow

• Clone the repo and run ./setup.sh
• Select the Sentry project and answer setup prompts
• When a Sentry alert fires, the AI reads the alert and repository context
• The AI writes a root‑cause sketch and creates a GitHub issue
• If the same error occurs again, the AI updates the existing issue instead of creating a duplicate
• Each run’s prompt and transcript are stored as a CI artifact

Evidence

An alert fires. Someone opens the stack trace, guesses, and closes the tab. Repeat tomorrow.

Source

A Sentry alert fires → the AI reads it against your repo and writes a root‑cause sketch

Source

That sketch is filed as a GitHub issue your team can actually act on

Source

The same error comes back → it updates the existing issue instead of filing a duplicate

Source

Every run's prompt and transcript is saved as a CI artifact, so you can check its reasoning

Source

It never resolves or ignores Sentry issues, and never closes the GitHub issue it opened.

Source

~5 minutes. No server, no database, free.

Source