Skip to main content

Source: ocean/docs/sentry-alerts-setup.md | ✏️ Edit on GitHub

Sentry Alerts & Issue Ownership Setup Guide

Overview​

This guide helps you configure alerts and issue ownership in your Sentry dashboard to maximize the value of error tracking.

1. Setting Up Alerts​

  1. Go to your Sentry project: https://sentry.io/organizations/goldfish-io/alerts/
  2. Click "Create Alert"

Critical Error Alert​

Type: Issue Alert

When: An issue is first seen
If: The issue is unhandled AND level is error
Then: Send email to engineering@goldfish.io

High Error Rate Alert​

Type: Metric Alert

When: Error rate is above 5% for 5 minutes
Environment: production
Then: Send Slack notification to #alerts channel

Performance Degradation Alert​

When: p95 transaction duration > 3 seconds
For: user-flow.* transactions
Environment: production
Then: Send email to engineering@goldfish.io

Failed Cron Job Alert​

When: Cron monitor weekly-build-health-check fails
Then: Send Slack notification to #devops channel

Authentication Failures Spike​

When: Event count > 50 in 10 minutes
Filter: transaction = "user-flow.authentication" AND status = "internal_error"
Then: Send PagerDuty alert

2. Issue Ownership Rules​

  1. Go to Settings → Projects → ocean → Issue Owners
  2. Click "Add Rule"

Ownership Rules by Path​

# Frontend Routes
path:src/routes/login.tsx                    #team-frontend @jane.doe
path:src/routes/dashboard/*                  #team-frontend @john.smith
path:src/components/ui/*                     #team-design @design.lead

# Backend Services
path:supabase/functions/handle-stripe-*      #team-billing @billing.lead
path:supabase/functions/graphql-v2/*         #team-backend @backend.lead

# Authentication & Security
path:src/services/auth.ts                    #team-security @security.lead
path:src/lib/jwt-verifier.ts                 #team-security @security.lead

# Performance Monitoring
transaction:user-flow.billing                #team-billing
transaction:user-flow.authentication         #team-security
transaction:user-flow.onboarding            #team-product

Ownership Rules by Tags​

# By User Plan
tags.user.plan:pro                          #team-customer-success
tags.user.plan:enterprise                   #team-enterprise @account.manager

# By Feature Area
tags.feature:billing                        #team-billing
tags.feature:authentication                 #team-security

# By Error Type
level:error url:*stripe.com*                #team-billing
level:error message:"GraphQL Error"         #team-backend

3. Slack Integration​

Setup Slack Integration​

  1. Go to Settings → Integrations → Slack
  2. Click "Add Workspace"
  3. Authorize Sentry

Configure Alert Channels​

#alerts          - All production errors
#alerts-billing  - Billing/Stripe issues
#alerts-security - Auth failures, suspicious activity
#alerts-perf     - Performance degradation

4. Alert Fatigue Prevention​

Smart Alerting Strategies​

  1. Use Alert Conditions

    Only alert if:
    - Error affects > 10 unique users
    - Error occurred > 100 times in 1 hour
    - Error is new (first seen)

  2. Time-based Suppression

    - Mute alerts during deployment windows
    - Different thresholds for business hours vs off-hours

  3. Grouping Similar Issues

    - Use fingerprinting to group related errors
    - Aggregate by transaction or URL pattern

5. Dashboard Configuration​

Create Custom Dashboards​

  1. Executive Dashboard

    • Error rate trend
    • User impact (affected users)
    • Top errors by frequency
    • Performance overview

  2. Developer Dashboard

    • Recent errors by owner
    • Unresolved issues assigned to me
    • Performance by transaction
    • Release health

  3. On-Call Dashboard

    • Real-time error feed
    • Critical alerts status
    • System health indicators
    • Recent deployments

6. Best Practices​

  1. Start Conservative

    • Begin with fewer, high-value alerts
    • Add more as you understand patterns

  2. Regular Review

    • Weekly review of alert effectiveness
    • Adjust thresholds based on false positives
    • Archive resolved ownership rules

  3. Clear Escalation

    • Define who gets alerted when
    • Have backup contacts for critical alerts
    • Document response procedures

  4. Use Alert Budgets

    • Max 5 alerts per developer per day
    • Critical alerts should be < 1 per week
    • Review if exceeding budgets

7. Monitoring Your Monitors​

Set up meta-alerts to track:

  • Alert volume (too many/too few)
  • Response time to alerts
  • False positive rate
  • Unassigned issues

Example Alert Workflow​

  1. Error Occurs → Sentry captures
  2. Ownership Rule → Assigns to team/person
  3. Alert Triggered → Notification sent
  4. Developer Responds → Acknowledges in Sentry
  5. Fix Deployed → Linked to release
  6. Alert Resolved → Automatically closed

Remember: Good alerting finds the balance between visibility and noise!