Auto-create and triage issues in Linear for your worst bugs
Set up an automated pipeline from error detection to AI-powered root cause analysis. Sentry catches the bug, creates a Linear issue via an alert rule, and the Sentry agent finds the root cause and suggests a code fix — all without leaving Linear.
Before you start
Accounts & access
- Sentry account with at least one project receiving errors
- Linear workspace with admin permissions to install integrations
- Sentry organization owner, manager, or admin role (required to install the Linear integration)
Tools
- sentry-cli installed (used to send a test event — no SDK changes required)
Knowledge
- Basic familiarity with Sentry alert rules
- Comfortable triggering a test error in your application
1 Install the Linear integration in Sentry
This is the first of two integrations you need. In Sentry, navigate to Settings → Integrations and search for Linear. Click Install and authenticate with your Linear workspace. This integration lets you create Linear issues from Sentry (manually or via alert rules), sync issue status bidirectionally, and link errors across both platforms. Once installed, you will see Linear listed under your active integrations with a green status.
Linear integration docs 2 Install and configure the Sentry Agent for Linear
This is the second integration — and the one that powers AI analysis. In Linear, go to Settings → Integrations → Sentry and install the Sentry Agent. During configuration, you will sign in to Sentry via OAuth to connect your account. Once installed, you can assign the Sentry agent to any Linear issue that has a linked Sentry issue, and it will automatically run Root Cause Analysis via Seer. Both integrations are required — the Linear integration (step 1) handles issue creation and sync, while the Linear Agent integration handles AI-powered analysis.
Sentry Agent for Linear docs 3 Create an alert rule that targets your worst bugs
Not every Sentry issue needs a Linear ticket. The key is filtering for high-impact errors so your backlog stays useful. In Sentry, go to Alerts → Create Alert Rule. For this walkthrough, set the condition to "A new issue is created" and add a filter for level:fatal. Under Actions, select "Create a Linear issue" and configure the target team, label, and priority. For production use, consider tighter filters to reduce noise:
- "An issue is seen more than 10 times in 1 hour" — catches errors that affect multiple users
- Filter by error level (
fatalorerroronly, notwarning) — skips non-critical noise - Filter by tag or transaction (e.g., only errors in your checkout flow) — targets what matters most
4 Trigger a test error
Send a test event to Sentry using sentry-cli to validate the pipeline end to end. This works regardless of what language or SDK your project uses. The --level fatal flag ensures it matches the alert filter from step 3. Within seconds, your alert rule should fire and create a Linear issue automatically.
sentry-cli send-event \
-m "Test: KeyError in process_order" \
--level fatal
5 Watch the Linear issue appear automatically
Within seconds of the error hitting Sentry, your alert rule fires and creates a Linear issue assigned to the team you configured in step 3. Open Linear and find the new issue — it includes a direct link back to the Sentry issue with the full stack trace, breadcrumbs, and context. No manual triage required. You can also check the Alert Activity page in Sentry to confirm the rule fired.
Alert notifications
6 Assign the Sentry agent to the Linear issue
Open the Linear issue that was just created and assign the Sentry agent to it. Because your alert rule auto-created this ticket, it already has a linked Sentry issue — which is what the agent needs to run. Once assigned, Seer kicks off automatically and performs Root Cause Analysis: it identifies the exact lines of code causing the error, explains why it fails, and suggests a code fix — all posted directly in the Linear issue. On your first run, the agent will ask whether you want it to automatically create pull requests for fixes going forward.
Learn about Seer
7 Apply the fix and verify the error stops
If you opted into automatic PR creation, the Sentry agent may have already opened a pull request with the fix. Otherwise, take the suggested code fix from the agent analysis and apply it to your codebase manually. Once deployed, check Issues in Sentry to confirm the error has stopped reproducing. When you mark the Linear issue as Done, the linked Sentry issue resolves automatically via bidirectional sync. The loop is closed — from error detection to triage to fix to resolution, all flowing through Linear.
That's it.
Your bugs now fix themselves (almost).
Sentry caught the error, your alert rule triaged it into Linear, and the Sentry agent diagnosed it with a root cause and code fix. Your debugging loop went from hours to minutes.
- Installed both the Linear integration (for issue sync) and the Linear Agent integration (for AI-powered analysis) in Sentry
- Created an alert rule that auto-creates Linear issues from Sentry errors
- Triggered a real error and watched it flow into Linear automatically
- Assigned the Sentry agent to a Linear issue to kick off Root Cause Analysis and get a code fix
- Optionally automated the entire flow using Linear triage rules
Pro tips
- 💡 Use Linear triage rules to fully automate the pipeline. Set a triage rule that automatically assigns the Sentry agent to any incoming issue from Sentry — no manual assignment needed.
- 💡 Set your alert rule conditions to target high-impact errors (e.g., 'seen by more than 10 users in 1 hour') so your Linear backlog does not fill with noise.
- 💡 Use Linear's label and priority fields in the alert action to auto-triage issues — for example, set P1 priority for errors affecting checkout flows.
- 💡 Seer works best when your GitHub repos are connected. The analysis is significantly better with access to your actual codebase alongside Sentry telemetry.
Common pitfalls
- ⚠️ You need both integrations installed. The Linear integration (installed in Sentry) handles issue creation and sync. The Linear Agent integration (installed in Linear) powers AI analysis. If you only install one, the pipeline will not work end to end.
- ⚠️ Alert rules only fire on new issue events by default. If you want to catch regressions of previously resolved issues, enable the 'A previously resolved issue has regressed' condition.
- ⚠️ Root Cause Analysis relies on having good stack traces. If your errors are minified or missing source maps, set up source maps first for better results.
- ⚠️ The Linear integration does not support private Linear teams. Make sure the target team is accessible to the integration.
Frequently asked questions
What's next?
Fix it, don't just track it.
Get started with the only application monitoring platform that gives your team AI-powered debugging directly in your project management workflow.