backup-recovery-policy

# Backup and Recovery Policy Skill ## Purpose This skill provides systematic guidance for implementing business continuity and disaster recovery within the CIA platform, ensuring data protection aligns with business impact analysis, RTO/RPO targets, and ISO 27001 Annex A.17 requirements. ## When to Use This Skill Apply this skill when: - ✅ Designing backup strategies for new systems or data stores - ✅ Configuring AWS backup services (RDS snapshots, S3 versioning, EBS backups) - ✅ Implementing disaster recovery procedures - ✅ Conducting quarterly backup restore tests - ✅ Defining RTO (Recovery Time Objective) and RPO (Recovery Point Objective) targets - ✅ Planning for data retention and archival - ✅ Responding to data loss incidents - ✅ Conducting business continuity planning Do NOT skip for: - ❌ Development/test environments (may contain production data copies) - ❌ Temporary data stores (may become permanent) - ❌ "Easily reproducible" data (validate recovery procedures) - ❌ Third-party managed services (verify backup capabilities) ## Business Impact-Driven Backup Framework ### RTO/RPO Targets by Classification | Classification | Business Impact | RPO Target | RTO Target | Backup Frequency | Retention | |----------------|-----------------|------------|------------|------------------|-----------| | **RESTRICTED** | Extreme | <15 minutes | <1 hour | Continuous replication | 30 days minimum | | **CONFIDENTIAL** | Very High | <4 hours | <4 hours | Hourly | 7 years (financial data) | | **INTERNAL** | Moderate | <24 hours | <24 hours | Daily | 3 years | | **PUBLIC** | Low | >24 hours | >72 hours | Weekly | Indefinite | ### Backup Strategy Decision Tree ```mermaid graph TD START["💾 Backup Need"] --> CLASSIFY{"🏷️ Data Classification?"} CLASSIFY -->|RESTRICTED| CRITICAL["🔴 Critical Backup<br/>Continuous Replication"] CLASSIFY -->|CONFIDENTIAL| HIGH["🟠 High Priority<br/>Hourly Backups"] CLASSIFY -->|INTERNAL| MEDIUM["🟡 Medium Priority<br/>Daily Backups"] CLASSIFY -->|PUBLIC| STANDARD["🟢 Standard<br/>Weekly Backups"] CRITICAL --> IMPACT{"📊 Business Impact?"} HIGH --> IMPACT MEDIUM --> IMPACT STANDARD --> IMPACT IMPACT -->|Financial System| FINANCE["💰 Financial Data<br/>7-year retention"] IMPACT -->|Core Operations| CORE["🏗️ Core Systems<br/>1-year retention"] IMPACT -->|Support Functions| SUPPORT["🛠️ Support Systems<br/>3-month retention"] FINANCE --> METHOD{"🔧 Backup Method?"} CORE --> METHOD SUPPORT --> METHOD METHOD -->|Database| RDS["📊 AWS RDS<br/>Automated Snapshots"] METHOD -->|Files| S3["📁 AWS S3<br/>Versioning + Lifecycle"] METHOD -->|Infrastructure| IaC["🏗️ Infrastructure as Code<br/>Git + CloudFormation"] RDS --> TEST["🧪 Quarterly Restore Test"] S3 --> TEST IaC --> TEST TEST --> MONITOR["📈 Monitoring & Alerts"] style CRITICAL fill:#D32F2F style HIGH fill:#FF9800 style MEDIUM fill:#FDD835 style STANDARD fill:#4CAF50 ``` ## Backup Strategies ### Full Backup Strategy **Definition**: Complete copy of all data at a point in time. **Use Cases**: - Initial backup establishment - Monthly comprehensive backups - Before major system changes - Regulatory compliance requirements **AWS Implementation**: ```yaml # CloudFormation template for full database backup Resources: DatabaseFullBackupFunction: Type: AWS::Lambda::Function Properties: FunctionName: cia-database-full-backup Runtime: python3.12 Handler: index.lambda_handler Role: !GetAtt BackupFunctionRole.Arn Timeout: 900 # 15 minutes Environment: Variables: RDS_INSTANCE_ID: !Ref CIADatabase BACKUP_BUCKET: !Ref BackupBucket Code: ZipFile: | import boto3 import datetime import os rds = boto3.client('rds') def lambda_handler(event, context): """ Create full RDS snapshot """ instance_id = os.environ['RDS_INSTANCE_ID'] timestamp = datetime.datetime.now().strftime('%Y%m%d-%H%M%S') snapshot_id = f"{instance_id}-full-{timestamp}" # Create snapshot response = rds.create_db_snapshot( DBSnapshotIdentifier=snapshot_id, DBInstanceIdentifier=instance_id, Tags=[ {'Key': 'BackupType', 'Value': 'Full'}, {'Key': 'CreatedBy', 'Value': 'Automated'}, {'Key': 'Retention', 'Value': '30days'} ] ) print(f"Full backup initiated: {snapshot_id}") return { 'statusCode': 200, 'body': snapshot_id } # Schedule monthly full backups FullBackupSchedule: Type: AWS::Events::Rule Properties: Name: cia-monthly-full-backup Description: Monthly full database backup ScheduleExpression: cron(0 2 1 * ? *) # 1st of month at 2 AM UTC State: ENABLED Targets: - Arn: !GetAtt DatabaseFullBackupFunction.Arn Id: FullBackupTarget ``` ## RTO and RPO Implementation ### Recovery Time Objective (RTO) **Definition**: Maximum acceptable time to restore service after an outage. **RTO Tiers**: | RTO Level | Time Window | Business Function Example | Implementation | |-----------|-------------|---------------------------|----------------| | **Instant** | <5 minutes | Financial transactions | Multi-AZ failover | | **Critical** | 5-60 minutes | Core database | Automated failover | | **High** | 1-4 hours | Application services | Blue-green deployment | | **Medium** | 4-24 hours | Analytics systems | Manual restore from backup | | **Standard** | >24 hours | Historical archives | Restore on demand | **RTO Configuration Example**: ```yaml # Multi-AZ RDS for instant failover (RTO <5 minutes) CIADatabase: Type: AWS::RDS::DBInstance Properties: DBInstanceIdentifier: cia-production-db Engine: postgres EngineVersion: "18.3" DBInstanceClass: db.t3.medium AllocatedStorage: 100 StorageType: gp3 StorageEncrypted: true KmsKeyId: !Ref DatabaseEncryptionKey # Multi-AZ for high availability (automatic failover) MultiAZ: true # Automated backups for point-in-time recovery BackupRetentionPeriod: 35 PreferredBackupWindow: "03:00-04:00" # Deletion protection DeletionProtection: true Tags: - Key: RTO Value: Critical-5to60min - Key: RPO Value: NearRealtime-1to15min - Key: BusinessImpact Value: VeryHigh ``` ### Recovery Point Objective (RPO) **Definition**: Maximum acceptable data loss measured in time. **RPO Tiers**: | RPO Level | Data Loss Window | Business Function Example | Backup Frequency | |-----------|------------------|---------------------------|------------------| | **Zero Loss** | <1 minute | Financial records | Synchronous replication | | **Near Real-time** | 1-15 minutes | Core database | Continuous backup | | **Minimal** | 15-60 minutes | Application data | 15-minute snapshots | | **Hourly** | 1-4 hours | User activity logs | Hourly backups | | **Daily** | 4-24 hours | Analytics data | Daily backups | | **Extended** | >24 hours | Archived data | Weekly backups | ## Quarterly Restore Testing ### Restore Test Procedure **Objective**: Verify backup integrity and validate RTO/RPO targets. **Frequency**: Quarterly minimum (ISO 27001 A.17.1.3) **Test Checklist**: - [ ] Select representative backup (full + incrementals) - [ ] Restore to isolated test environment - [ ] Verify data integrity (checksums, row counts) - [ ] Test application functionality against restored data - [ ] Measure actual recovery time vs RTO target - [ ] Measure data loss vs RPO target - [ ] Document results and lessons learned - [ ] Update recovery procedures if needed **Automated Restore Test Script**: ```bash #!/bin/bash # Quarterly backup restore test # Tests RTO/RPO compliance and backup integrity set -euo pipefail TEST_DATE=$(date +%Y%m%d-%H%M%S) TEST_REPORT="backup-restore-test-${TEST_DATE}.md" TEST_INSTANCE="cia-restore-test-${TEST_DATE}" log() { echo "[$(date -u +"%Y-%m-%d %H:%M:%S UTC")] $*" | tee -a "${TEST_REPORT}" } # Start restore test log "# Quarterly Backup Restore Test" log "" log "**Test Date**: $(date -u +"%Y-%m-%d %H:%M:%S UTC")" log "**Tester**: CEO" log "**Test Instance**: ${TEST_INSTANCE}" log "" # Step 1: Identify latest backup log "## Step 1: Identify Latest Backup" SNAPSHOT_ID=$(aws rds describe-db-snapshots \ --db-instance-identifier cia-production-db \ --query 'DBSnapshots | sort_by(@, &SnapshotCreateTime) | [-1].DBSnapshotIdentifier' \ --output text) log "- Latest snapshot: ${SNAPSHOT_ID}" SNAPSHOT_TIME=$(aws rds describe-db-snapshots \ --db-snapshot-identifier "${SNAPSHOT_ID}" \ --query 'DBSnapshots[0].SnapshotCreateTime' \ --output text) log "- Snapshot time: ${SNAPSHOT_TIME}" log "" # Step 2: Restore snapshot to test instance log "## Step 2: Restore Database" START_TIME=$(date +%s) log "- Initiating restore..." aws rds restore-db-instance-from-db-snapshot \ --db-instance-identifier "${TEST_INSTANCE}" \ --db-snapshot-identifier "${SNAPSHOT_ID}" \ --db-instance-class db.t3.small \ --publicly-accessible false \ --no-multi-az \ --tags Key=Purpose,Value=RestoreTest Key=TestDate,Value="${TEST_DATE}" # Wait for instance to be available log "- Waiting for instance to become available..." aws rds wait db-instance-available --db-instance-identifier "${TEST_INSTANCE}" END_TIME=$(date +%s) RESTORE_DURATION=$((END_TIME - START_TIME)) log "- ✅ Restore completed in ${RESTORE_DURATION} seconds" log "" # Step 3: Verify data integrity log "## Step 3: Data Integrity Verification" # Get endpoint DB_ENDPOINT=$(aws rds describe-db-instances \ --db-instance-identifier "${TEST_INSTANCE}" \ --query 'DBInstances[0].Endpoint.Address' \ --output text) log "- Database endpoint: ${DB_ENDPOINT}" # Verify row counts log "- Verifying table row counts..." psql -h "${DB_ENDPOINT}" -U cia_user -d cia_database -c "\ SELECT schemaname, tablename, n_live_tup as row_count \ FROM pg_stat_user_tables \ ORDER BY n_live_tup DESC \ LIMIT 10;" | tee -a "${TEST_REPORT}" # Step 4: Cleanup log "## Step 4: Cleanup" log "- Deleting test instance..." aws rds delete-db-instance \ --db-instance-identifier "${TEST_INSTANCE}" \ --skip-final-snapshot \ --delete-automated-backups log "- ✅ Test instance cleanup initiated" log "" # Summary log "## Test Summary" log "" log "| Metric | Target | Actual | Status |" log "|--------|--------|--------|--------|" log "| RTO | <4 hours | $(($RESTORE_DURATION / 60)) minutes | ✅ Pass |" log "| Data Integrity | 100% | Verified | ✅ Pass |" log "" echo "✅ Restore test completed. Report: ${TEST_REPORT}" ``` ## Retention Policies ### Retention by Classification | Classification | Retention Period | Rationale | Disposal Method |