sentry-alerts-issues（哨兵警报问题）

Sentry - Alerts & Issue Management

Configure alerts, manage issues, and set up notifications.

Alert Types

Issue Alerts

Trigger when an issue matches specific conditions:

• New issue created
• Issue changes state (resolved → unresolved)
• Issue frequency exceeds threshold
• Issue affects specific users/releases

Metric Alerts

Trigger based on aggregate metrics:

• Error count exceeds threshold
• Transaction latency (p50, p95, p99)
• Failure rate percentage
• Apdex score drops

Uptime Monitoring

Monitor endpoint availability:

• HTTP status codes
• Response time thresholds
• SSL certificate expiration

Alert Conditions

Frequency-Based

When an issue is seen more than 100 times in 1 hour

User Impact

When an issue affects more than 50 unique users in 1 hour

First Seen

When a new issue is created

Regression

When an issue changes state from resolved to unresolved

Alert Actions

Notification Channels

• Email: Individual or team distribution
• Slack: Channel or direct messages
• PagerDuty: On-call escalation
• Discord: Webhook notifications
• Microsoft Teams: Channel posts
• Webhooks: Custom integrations

Issue Actions

• Assign to team member
• Add tags
• Create ticket (Jira, Linear, GitHub)

Issue States

Lifecycle

1. Unresolved: Active issue needing attention
2. Resolved: Fixed in a release
3. Ignored: Intentionally dismissed
4. Archived: Historical, no longer tracked

Substates

• For Review: Needs triage
• Escalating: Increasing in frequency
• Regressed: Previously resolved, now recurring
• New: First seen recently

Issue Management

Merge Issues

Combine duplicate issues with different stack traces:

1. Select issues to merge
2. Choose primary issue
3. All events appear under primary

Ignore Rules

Ignore this issue:
- Forever
- Until it happens again
- For the next 24 hours
- Until it affects 100 users

Ownership Rules

# Define in project settings
path:src/payments/* #payments-team
path:src/auth/* user@example.com
tags.component:checkout #checkout-team

API Integration

List Issues

curl "https://sentry.io/api/0/projects/{org}/{project}/issues/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN"

Update Issue

curl -X PUT "https://sentry.io/api/0/issues/{issue_id}/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"status": "resolved"}'

Create Alert Rule

curl -X POST "https://sentry.io/api/0/projects/{org}/{project}/rules/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "High Error Rate",
    "conditions": [...],
    "actions": [...],
    "frequency": 30
  }'

Integration Examples

Slack Alert

{
  "action": "notify_slack",
  "workspace": "your-workspace",
  "channel": "#alerts",
  "tags": ["level", "environment"]
}

PagerDuty Escalation

{
  "action": "notify_pagerduty",
  "service": "your-service-key",
  "severity": "critical"
}

Jira Ticket Creation

{
  "action": "create_jira_ticket",
  "integration": "jira-integration-id",
  "project": "PROJ",
  "issueType": "Bug"
}

Best Practices

1. Start with conservative thresholds, tune over time
2. Use different channels for different severities
3. Set up on-call rotation for critical alerts
4. Review and archive stale issues regularly
5. Define ownership rules for automatic assignment
6. Create runbooks for common alert types
7. Use metric alerts for SLO monitoring