Backup window appropriate for RPO

Severity: Recommended

Backup timing should be intentional: during low-traffic periods to minimize performance impact, and frequent enough to meet business RPO requirements.

Check automatically:

1. Check backup window timing:

# AWS RDS backup window (UTC)
aws rds describe-db-instances --query "DBInstances[].{ID:DBInstanceIdentifier,BackupWindow:PreferredBackupWindow}" --output table

# GCP Cloud SQL backup start time
gcloud sql instances describe INSTANCE_NAME --format="get(settings.backupConfiguration.startTime)"

2. Check Terraform for backup window:

# AWS RDS
grep -rE "preferred_backup_window|backup_window" --include="*.tf" 2>/dev/null

# GCP Cloud SQL
grep -rE "start_time.*backup" --include="*.tf" 2>/dev/null

3. Check cron schedules for scripted backups:

# Look for backup cron patterns
grep -rE "cron|schedule" --include="*.yml" --include="*.yaml" --include="*.sh" 2>/dev/null | grep -iE "backup|dump"

# Check Kubernetes CronJobs
kubectl get cronjobs -A 2>/dev/null | grep -iE "backup|dump"

4. Check backup frequency:

# AWS RDS - automated backups are daily, but PITR provides continuous
# Check snapshot frequency for manual snapshots
aws rds describe-db-snapshots --snapshot-type manual --query "DBSnapshots[].SnapshotCreateTime" --output text | head -10

# GCP - check backup frequency
gcloud sql instances describe INSTANCE_NAME --format="get(settings.backupConfiguration.transactionLogRetentionDays)"

Ask user:

• "When do your backups run? Is this during low-traffic periods?"
• "What's your RPO (Recovery Point Objective) - how much data loss is acceptable?"
• "Does your backup frequency match your RPO?"
• "Have you noticed performance impact during backup windows?"

RPO considerations:

• If RPO is 1 hour, daily backups aren't enough (need PITR or hourly snapshots)
• If RPO is 24 hours, daily backups are sufficient
• PITR with continuous WAL archiving effectively gives RPO of seconds/minutes

Cross-reference with:

• HA-003 (backups exist - this item is about timing)
• HA-005 (PITR - if enabled, provides continuous protection regardless of window)
• Section 34 (Rollback & Recovery - RPO/RTO definitions)
• MON-002 (database performance - backup impact on queries)

Pass criteria:

• Backup window defined and intentional (not just default)
• Window is during low-traffic period for the application
• Backup frequency aligns with business RPO requirements
• No significant performance degradation during backups

Fail criteria:

• Default backup window never reviewed
• Backups run during peak traffic causing performance issues
• RPO requirement is 1 hour but backups are daily (and no PITR)
• Backup window conflicts with other maintenance

Evidence to capture:

• Backup window (time in UTC and local timezone)
• Backup frequency (daily, hourly, continuous)
• Business RPO requirement
• Whether PITR fills the gap between snapshots
• Any known performance impact

---