incident-runbook-generator（事件运行手册生成器）

Incident Runbook Generator

Create actionable runbooks for common incidents.

Runbook Template

# Runbook: Database Connection Pool Exhausted

**Severity:** P1 (Critical)
**Estimated Time to Resolve:** 15-30 minutes
**Owner:** Database Team (On-call)

## Symptoms

- Application errors: "connection pool exhausted"
- Increased API latency (>5s)
- Failed health checks
- CloudWatch alarm: `DatabaseConnectionsHigh`

## Detection

- Alert: DatabaseConnectionPoolExhausted
- Metrics: `active_connections > max_connections * 0.9`
- Logs: "Error: connect ETIMEDOUT"

## Immediate Actions (5 min)

1. **Verify the issue**

# Check current connections SELECT count(*) FROM pg_stat_activity;

1. Identify long-running queries SELECT pid, now() - pg_stat_activity.query_start AS duration, query FROM pg_stat_activity WHERE state = 'active' ORDER BY duration DESC LIMIT 10;
2. Kill blocking queries (if safe) SELECT pg_terminate_backend(pid) FROM pg_stat_activity WHERE state = 'idle in transaction' AND now() - state_change > interval '5 minutes';

Mitigation (10 min)

1. Scale up connection pool (temporary) # Update RDS parameter group aws rds modify-db-parameter-group \ --db-parameter-group-name prod-params \ --parameters "ParameterName=max_connections,ParameterValue=200"
2. Restart application (if needed) kubectl rollout restart deployment/api
3. Monitor recovery watch -n 5 'psql -c "SELECT count(*) FROM pg_stat_activity;"'

Root Cause Investigation

Check for:

• Recent deployment (new code with connection leaks)
• Traffic spike (legitimate or DDoS)
• Slow queries holding connections
• Connection pool configuration too small
• Application not releasing connections

Rollback Steps

If caused by deployment:

# Rollback to previous version
kubectl rollout undo deployment/api

# Verify
kubectl rollout status deployment/api

Communication Template

Initial (within 5 min):

🚨 INCIDENT: Database connection pool exhausted
Status: Investigating
Impact: API errors and slowness
ETA: 15-30 min
Next update: 10 min

Update (every 10 min):

UPDATE: Killed long-running queries
Status: Mitigating
Impact: Still degraded, improving
Actions: Scaling connection pool
Next update: 10 min

Resolution:

✅ RESOLVED: Database connections normalized
Duration: 25 minutes
Root cause: Connection leak in v2.3.4
Fix: Rolled back to v2.3.3
Follow-up: Bug fix PR #1234
Postmortem: [link]

Prevention

• Add connection pool metrics to dashboards
• Implement connection timeout (30s)
• Add connection leak detection in tests
• Set up pre-deployment load testing
• Review connection pool sizing

Related Runbooks

• Database High CPU
• Slow Database Queries
• Application OOM

## Output Checklist

- [ ] Symptoms documented
- [ ] Detection criteria
- [ ] Step-by-step actions
- [ ] Owner assigned
- [ ] Rollback procedure
- [ ] Communication templates
- [ ] Prevention measures
ENDFILE