Skip to main content

Overview

Connect Sentry to Continue Mission Control to enable agents to automatically detect, analyze, and resolve production errors. When Sentry is enabled, Continue can generate PRs to fix issues, analyze error patterns, and maintain application health.

What You Can Do with Sentry Integration

  • Automatically generate PRs to resolve new errors
  • Analyze error patterns and identify root causes
  • Prioritize issues based on impact and frequency
  • Create detailed bug reports with stack traces
  • Monitor error trends across releases

Setup

1

Navigate to Integrations

2

Connect Sentry

Click “Connect” and select Sentry. You’ll need the following credentials:
  • Sentry Organization Slug: Your organization name (e.g., “my-company”)
  • Auth Token: Internal integration token from Sentry
  • Client Secret: For webhook signature verification
3

Configure Sentry Internal Integration

In your Sentry account:
  1. Create a new internal integration
  2. Set the webhook URL to https://api.continue.dev/webhooks/sentry
  3. Select “Read and Write” access for “Issue and Event”
  4. Select “issue” in the webhooks section and click Save Changes
  5. Copy the Auth Token and Client Secret
4

Add Credentials to Mission Control

Paste all three values into the integration form and click “Create Connection”
How to get these credentials:
  1. Create a new internal integration in Sentry
  2. Set the webhook URL to https://api.continue.dev/webhooks/sentry
  3. Select “Read and Write” access for “Issue and Event”
  4. Select “issue” in the webhooks section and click Save Changes
  5. Find your organization slug in the Sentry URL (your-slug.sentry.io)
  6. Copy the Auth Token and Client Secret from the integration

Running Sentry Agents in Mission Control

You can run Sentry-connected agents in two ways:

1. Manual Tasks

Trigger agents on-demand for error analysis:
  1. Go to Mission Control Agents
  2. Select or create a Sentry-enabled agent
  3. Click “Run Agent” and provide your task description
  4. Monitor progress and review results in real-time
Example Tasks:
  • “Analyze the top 10 errors from the last 24 hours”
  • “Create a PR to fix the authentication timeout error”
  • “Generate a report on errors affecting mobile users”

2. Automated Workflows

Set up agents to run automatically:
  • Webhook-triggered: Execute when new Sentry errors occur
  • Scheduled: Run daily or weekly error analysis
  • Threshold-based: Trigger when error rates exceed limits
Start with manual tasks to refine your prompts, then convert successful workflows to automations for continuous error monitoring.

Integration with GitHub

Combine Sentry with GitHub integration for a complete workflow:
1

Connect Both Integrations

Enable both Sentry and GitHub integrations in Mission Control
2

Create a Unified Agent

Build an agent that:
  • Receives Sentry error notifications
  • Analyzes the error and finds the problematic code
  • Creates a PR with a fix
  • Adds the PR link to the Sentry issue
3

Set Up Automated Workflow

Configure the agent to run automatically on new Sentry issues

Monitoring Agent Activity

Track your agent’s error resolution performance:
  1. View in Mission Control: See all agent runs and their outcomes
  2. Check Sentry: Verify that issues are being resolved
  3. Review PRs: Ensure quality of generated fixes
  4. Monitor Metrics: Track resolution time and success rate

Troubleshooting

Problem: Agent isn’t running when new errors occurSolutions:
  • Verify webhook URL is correct in Sentry
  • Check that “issue” is selected in webhook events
  • Ensure Client Secret matches in both Sentry and Mission Control
  • Review webhook delivery logs in Sentry
Problem: Agent returns errors when trying to fetch issue detailsSolutions:
  • Verify Auth Token has “Read and Write” access for “Issue and Event”
  • Check that organization slug is correct
  • Ensure token hasn’t expired
  • Confirm project permissions in Sentry
Problem: Generated PRs don’t reference the Sentry issueSolutions:
  • Ensure both Sentry and GitHub integrations are connected
  • Verify agent has permission to access both services
  • Check that the agent prompt includes instructions to link issues
  • Review agent logs for errors

Support & Resources