NIST Cybersecurity Framework 2.0¶
Document Control:
Version: 2.0 Last Updated: April 2026 Owner: Paul Leone Framework Version: NIST CSF 2.0
Overview¶
Overall Maturity: Tier 3 (Repeatable), approaching Tier 4 (Adaptive). Version 2.0 extends scope to multi-cloud hybrid infrastructure spanning AWS (us-east-2), Azure (North Central US), and GCP (us-central1), connected via Tailscale WireGuard mesh VPN. All cloud nodes integrate with the on-premises management toolchain (Wazuh EDR, PatchMon, Checkmk, Ansible, Uptime Kuma).
GOVERN (GV) — Tier 3¶
GV.OC — Organizational Context¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| GV.OC-01 | Context of organization and security strategy established | Lab mission statement defines security-first architecture principles; documented security objectives aligned with learning goals; compliance requirements documented across NIST, CIS, ISO 27001, and PCI-DSS frameworks | Cloud scope included in mission documentation: multi-cloud hybrid architecture (AWS/Azure/GCP) added as a demonstrated capability alongside on-premises SOC operations |
| GV.OC-02 | Internal and external stakeholders understood | N/A (Personal lab) | — |
| GV.OC-03 | Legal/regulatory/contractual requirements understood | Conceptual compliance with NIST, CIS, ISO 27001, PCI-DSS frameworks; no actual regulatory obligations (personal lab) | Cloud provider terms of service acknowledged. No cardholder data processed on cloud nodes |
| GV.OC-04 | Critical objectives, capabilities, services understood | Core services documented with criticality tiers: Tier 1 (SIEM/EDR), Tier 2 (Firewalls/DNS/Identity), Tier 3 (Supporting services) | Cloud nodes classified as Tier 2. Cloud-native security services (GuardDuty, SCC, Defender for Cloud) classified as Tier 1 alongside Splunk and Wazuh |
| GV.OC-05 | Outcomes of cybersecurity strategy communicated | Security metrics tracked in Grafana dashboards (20+); quarterly reviews conducted against documented objectives | Cloud security posture included in dashboard scope: Checkmk cloud node health, PatchMon cloud patch compliance, Uptime Kuma cloud tunnel availability |
GV.RM — Risk Management Strategy¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| GV.RM-01 | Risk management objectives established | Risk-based vulnerability prioritization via CVSS scoring; patch management SLAs by severity: Critical <72h, High <7d; threat-informed architecture | Same CVSS-based SLAs applied to cloud node findings from PatchMon, AWS Inspector, GCP SCC, and Azure Defender for Cloud |
| GV.RM-02 | Risk appetite and tolerance defined | Critical CVEs: MTTR <72h; High CVEs: MTTR <7d; acceptable false positive rate <5%; 95% patch compliance target | Cloud node patch SLAs enforced identically to on-premises. Azure Spot instance eviction risk accepted with documented Deallocate policy and restart runbook |
GV.PO — Policy¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| GV.PO-01 | Policy establishes behavioral expectations | SSH hardening policy (CIS Benchmark); TLS 1.3 minimum policy; vulnerability remediation SLAs; comprehensive logging policy (100% security event coverage); policies version-controlled in Git | Cloud policies: no public management ports on any cloud instance; all SSH/RDP restricted to homelab subnets (192.168.0.0/16) via provider-native firewall rules; Tailscale WireGuard mandatory for all management paths; CIS SCA enforcement via Wazuh on all cloud nodes |
GV.SC — Cybersecurity Supply Chain Risk Management¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| GV.SC-02 | Suppliers and third parties included in risk management | Vetted open-source projects; trusted Docker Hub publishers; official OS repositories; GitHub security alerts monitored | AWS IAM instance profiles (least-privilege, no long-lived access keys on EC2); GCP service accounts with minimal IAM role assignments; Azure VMs provisioned via ARM template with securestring parameters; provider-native image signing verified |
| GV.SC-04 | Suppliers and partners routinely assessed | GitHub security alerts monitored; software update frequency tracked; planned CVE monitoring for dependencies | Cloud provider security posture assessed continuously: AWS Security Hub, GCP SCC security health analytics, Azure Defender for Cloud Secure Score. Provider SOC 2 and ISO 27001 certifications acknowledged |
| GV.SC-05 | Response and recovery planning for supply chain incidents | Snapshot-before-update strategy; rollback procedures documented | Cloud node recovery: ARM template (Azure) and Ansible playbooks enable rebuild from code; Tailscale re-enrollment and Wazuh agent reinstallation documented in runbooks; cloud node images (AMIs, GCP disk snapshots) available for rapid recovery |
IDENTIFY (ID) — Tier 4¶
ID.AM — Asset Management¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| ID.AM-01 | Inventories of hardware managed | Checkmk infrastructure monitoring; Prometheus node exporters (30+ hosts); Proxmox asset database; external Excel inventory | Cloud hardware inventory added: AWS EC2 (i-0204e2e26d7f30631 t3.micro Amazon Linux 2, i-03e9efec4fa3fe644 c7i-flex.large Windows Server 2025); GCP (2× e2.micro, us-central1-c); Azure (Standard_B2ats_v2 Ubuntu 24.04, Standard_E2s_v3 Windows Server 2025). All 8 cloud nodes enrolled in Checkmk, Wazuh, and PatchMon via Tailscale |
| ID.AM-02 | Inventories of software and applications managed | PatchMon tracks 5,000+ packages across 30+ hosts; WUD tracks 50+ containers; Nessus software inventory; Wazuh agent inventory (25+ endpoints) | PatchMon extended to 6 Linux cloud nodes with daily NVD CVE correlation. Windows cloud nodes (AWS, Azure) tracked via WSUS with automated approval workflows. Wazuh agents on all 8 cloud nodes provide real-time package and process inventory |
| ID.AM-03 | Network diagrams and communication flows documented | Network topology documented; Traefik routing architecture; VLAN segmentation diagrams; data flow maps maintained in Git | Cloud topology added: Tailscale mesh routing tables (9 enrolled nodes, tailf07c05.ts.net), per-provider VPC/VNet CIDR allocations (AWS 172.31.0.0/16, GCP 10.128.0.0/16, Azure 10.130.0.0/16), subnet router configurations, and cloud-to-on-premises routing tables documented |
| ID.AM-04 | External systems and network connections catalogued | VPN connections documented (Tailscale, PIA, Cloudflare Tunnels); external DNS resolvers tracked; internet egress points mapped | Cloud VPC/VNet connections catalogued: AWS VPC (vpc-07db6afa9c6097fe4), GCP default VPC (10.128.0.0/16, us-central1), Azure homelab-vnet2 (10.130.0.0/16, North Central US). Tailscale subnet routers advertising cloud CIDRs into homelab documented |
| ID.AM-05 | Resources prioritized by classification and criticality | Tier 1: SIEM, EDR, firewalls. Tier 2: DNS, identity management. Tier 3: Supporting services | Cloud nodes classified as Tier 2. Cloud-native security services (GuardDuty, SCC, Defender for Cloud, Sentinel) classified as Tier 1 alongside on-premises Splunk and Wazuh. Azure Spot instance (Standard_E2s_v3) noted as lower-availability Tier 2 with documented restart procedure |
| ID.AM-07 | Inventories of data and information managed | Sensitive data classification (logs, credentials, backups); data flow mapping; retention policies documented | Cloud data inventory: AWS S3 CloudTrail log buckets (SSE, 90-day lifecycle); GCP Cloud Storage (default encryption, audit logging); Azure Log Analytics workspace homelab-log (90-day retention). All 8 cloud nodes included in Wazuh FIM monitoring scope via Tailscale |
| ID.AM-08 | Systems, hardware, software, and services authorized | Authorized software list maintained; unauthorized application detection via Nessus/Wazuh; procurement approval process | Cloud resources authorized: AWS EC2 in us-east-2 under documented IAM roles; GCP Compute in us-central1-c under minimal service accounts; Azure VMs in North Central US provisioned via ARM template. All cloud nodes enrolled in Tailscale tailnet with ACL-enforced access policy |
ID.RA — Risk Assessment¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| ID.RA-01 | Vulnerabilities identified and documented | OpenVAS + Nessus vulnerability scanning; CVSS scoring; CVE correlation with NVD; TheHive vulnerability tracking | Wazuh CIS SCA on all 8 cloud nodes daily; PatchMon CVE correlation for 6 Linux cloud nodes; AWS Inspector; GCP SCC vulnerability assessment; Azure Defender for Cloud recommendations. All cloud findings visible in unified PatchMon dashboard |
| ID.RA-02 | Cyber threat intelligence received | MISP threat intelligence platform; CrowdSec community feeds; AlienVault OTX, abuse.ch, Emerging Threats; Shuffle threat aggregation | AWS GuardDuty ML-based behavioral detection (VPC flow logs, CloudTrail, DNS query logs); GCP Security Command Center asset and threat findings; Azure Sentinel KQL detection rules (NSG deny spikes, new-location sign-ins, domain join anomalies, AMA data gaps). Cloud findings forwarded to on-premises SIEM via Wazuh agents |
| ID.RA-04 | Potential impacts and likelihoods identified | Risk-based remediation prioritization; exploit availability assessed; CVSS temporal scoring; asset criticality factored | Cloud risk factors: cloud node public IPs used only for Tailscale UDP handshake assessed as low-risk. Azure Spot instance eviction risk documented with Deallocate policy. GCP e2.micro resource constraints factored into availability risk assessment |
| ID.RA-08 | Processes for receiving and analyzing vulnerability disclosures established | Shuffle vulnerability disclosure workflow; NIST NVD monitoring; vendor advisory tracking; CVE analysis pipeline | Cloud vulnerability disclosure: AWS Inspector findings routed to CloudWatch Events; GCP SCC findings generate alerting policy notifications; Azure Defender for Cloud recommendations surfaced in homelab-log and Sentinel. All cloud findings fed into TheHive for tracking against same MTTR SLAs as on-premises |
| ID.RA-10 | Critical suppliers and dependencies included in risk assessments | Planned SBOM tracking; dependency vulnerability scanning; third-party risk assessments | Cloud provider dependency risk: AWS, Azure, and GCP availability tracked via Uptime Kuma. Tailscale (WireGuard mesh VPN) identified as critical dependency for all cloud management access; dual on-premises pfSense nodes (primary and secondary Tailscale subnet routers) provide redundancy |
ID.IM — Improvement¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| ID.IM-01 | Improvements identified from detection and response activities | Vulnerability trends tracked in Grafana; CIS Benchmark compliance scores monitored; patch compliance dashboard; continuous security posture improvement | Cloud security posture improvements tracked: PatchMon cloud patch compliance trends; Checkmk cloud node health trending; AWS Security Hub score trending, GCP SCC findings trend, Azure Defender for Cloud Secure Score monitored |
| ID.IM-03 | Response and recovery plans incorporate lessons learned | Post-incident reviews drive playbook updates; TheHive lessons learned tracking; Shuffle workflow optimization based on execution data | Cloud incident lessons: Azure Spot instance eviction experience documented; Tailscale tunnel recovery procedure refined; cloud node Wazuh re-enrollment runbook tested and updated following agent version upgrades |
PROTECT (PR) — Tier 4¶
PR.AA — Identity Management, Authentication and Access Control¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| PR.AA-01 | Identities and credentials issued, managed, and verified | Authentik SSO integration; centralized SSH key management via Ansible; individual user accounts; MFA enforcement via Authentik TOTP | Linux cloud nodes use SSH Ed25519 keys managed via Ansible (same key distribution playbook as on-premises). AWS Windows Server 2025 and Azure Windows Server 2025 domain-joined to home.com via Tailscale — Kerberos authentication, Group Policy, and WSUS extended to cloud workloads |
| PR.AA-02 | Identities authenticated | SSH key-based authentication; Authentik SSO; MFA enforcement; certificate-based authentication (Step-CA) | SSH key-only access for all Linux cloud nodes (passwords disabled globally via Ansible). No public management ports — all SSH/RDP access via Tailscale with device-bound WireGuard authentication. Windows cloud nodes authenticate via domain Kerberos |
| PR.AA-05 | Access permissions and authorizations managed | Authentik RBAC groups; SSH sudo enforcement; Traefik middleware access control; least-privilege principle | AWS Security Groups (tailscale-access + lab-services) restrict all inbound management to homelab subnets and Tailnet CGNAT (100.64.0.0/10). GCP VPC firewall policy restricts SSH to 192.168.0.0/16. Azure NSG homelab-nsg2 explicit allow-list with DenyAllInBound at priority 65500. AWS IAM instance profiles scoped to minimum required actions; no wildcard permissions |
| PR.AA-06 | Physical access managed | Physical security (personal lab): locked server rack, limited access; environmental monitoring; backup power (UPS) | Cloud physical security delegated to provider (AWS, GCP, Azure maintain SOC 2 / ISO 27001 certified data centers) |
PR.DS — Data Security¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| PR.DS-01 | Data at rest protected | Encrypted backups (AES-256); TLS in transit; scan credential encryption; SSH private keys encrypted; encrypted log transmission (syslog-ng TLS); immutable SIEM indexes | AWS S3 CloudTrail log buckets with SSE; GCP persistent disks encrypted by default (Google-managed keys); Azure Managed Disks encrypted at rest by default; Azure Log Analytics homelab-log data encrypted at rest. All 8 cloud nodes included in Wazuh FIM monitoring scope |
| PR.DS-02 | Data in transit protected | TLS 1.3 encryption (Traefik); Ed25519 SSH keys; DNSSEC validation; encrypted VPN tunnels (Tailscale, WireGuard); IPSec IKEv2 AES-256-GCM site-to-site | All management traffic between cloud nodes and on-premises encrypted via Tailscale WireGuard (ChaCha20-Poly1305). SSH AES-256-GCM for Linux cloud node terminal access. Cloud VPC/VNet internal traffic uses provider-encrypted private networking. No plaintext management paths exist across any cloud provider |
| PR.DS-10 | Integrity and authenticity of hardware and software verified | Package signature verification; container image verification (SHA-256); Step-CA certificate validation | AWS AMI and GCP Compute image provenance tracked; Azure Marketplace image signatures verified. Ansible new_install_baseline_roles.yml validates package signatures on cloud nodes at provisioning time |
| PR.DS-11 | Data disposal practices established | Secure deletion procedures; backup rotation policies; log retention limits; expired certificate cleanup | AWS EC2 EBS volumes deleted on instance termination (documented in runbook). GCP persistent disk deletion policy documented. Azure Managed Disk cleanup included in decommissioning checklist. Cloud-native audit logs subject to 90-day retention policy |
PR.IR — Technology Infrastructure Resilience¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| PR.IR-01 | Networks and environments protected | HA DNS failover (dual Unbound and Technitium); Traefik zero-downtime reloads; firewall clustering; network segmentation | AWS VPC Security Groups enforce network-layer segmentation. GCP VPC firewall policy (homelab) with Google Threat Intelligence blocking — TOR exit nodes, known malicious IPs, sanctioned countries (CU, IR, KP, SY, Crimea, Donetsk/Luhansk). Azure NSG homelab-nsg2 default-deny. GCP Cloud Armor three-tier WAF/DDoS policies |
| PR.IR-02 | Security architectures established and maintained | Defense-in-depth architecture; zero-trust principles; documented security controls; continuous assessment | Tailscale ACL-enforced mesh (default deny, explicit allow policy) extends zero-trust posture to cloud nodes. Per-provider enclave gateway model: AWS Security Groups, GCP VPC firewall policy, Azure NSG. GCP Cloud Armor application-level WAF (SQLi/XSS blocking via v33-stable expression sets) |
| PR.IR-04 | Adequate capacity ensured | Prometheus capacity monitoring; Pulse hypervisor monitoring; disk space alerts; resource trending | AWS CloudWatch Agent (CPU, memory, disk, network per EC2 instance); GCP Cloud Monitoring Ops Agent for Compute nodes; Azure Monitor VM metric alert rules (7 active: Available Memory Bytes, Data Disk IOPS Consumed %, Network In/Out Total, OS Disk IOPS, Percentage CPU, VM Availability). All feed Checkmk on-premises dashboard and Discord alerting |
| PR.IR-05 | Continuity prioritized by business criticality | Service prioritization (Tier 1-3); documented recovery priorities; RTO/RPO defined | Tailscale subnet router redundancy via dual on-premises pfSense nodes. Azure Spot instance eviction handled by Deallocate policy with documented restart procedure. Cloud node RTO target <2hr via Ansible rebuild playbooks validated |
PR.PS — Platform Security¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| PR.PS-01 | Configuration management practices established | SSH config via Ansible; Traefik YAML in Git; DNS zones version-controlled; PatchMon configuration tracking; CIS Benchmark audits; Terraform IaC; configuration drift detection | Ansible linux_hardening.yml and new_install_baseline_roles.yml applied to all 4 Linux cloud nodes via SSH over Tailscale. Windows cloud nodes managed via GPO (domain-joined to home.com). Azure ARM template (azuredeploy.json) version-controlled in Git. All cloud configurations stored in Git |
| PR.PS-03 | Hardware and software maintained | Multi-platform patch management (PatchMon, WSUS, WUD, Watchtower); vulnerability remediation workflows; automated updates where appropriate | PatchMon daily checks on 3 Linux cloud nodes (Amazon Linux 2, Debian 13.4, Ubuntu 24.04 LTS). WSUS manages Windows cloud nodes (AWS, Azure Windows Server 2025) with automated approval workflows. Critical <72h MTTR SLA maintained across all cloud nodes |
| PR.PS-04 | Log records generated and managed | 100% security event logging to SIEM; 90-day retention; structured JSON format; comprehensive audit trails | AWS CloudTrail (all management events, delivered to S3 with SSE); GCP Admin Activity logs (always-on, cannot be disabled); Azure Activity Log and Monitor; all supplemented by Wazuh agent event forwarding from cloud nodes to on-premises Splunk. Cloud-native logs retained in provider storage with 90-day retention matching on-premises policy |
DETECT (DE) — Tier 3¶
DE.AE — Adverse Event Analysis¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| DE.AE-02 | Potentially adverse events analyzed | Vulnerability trending; exploit likelihood assessment; Cortex automated enrichment; MISP threat intelligence correlation; Shuffle orchestrated analysis workflows | AWS GuardDuty findings enriched with threat intelligence and routed to CloudWatch Events; GCP SCC findings generate alerting policy notifications; Azure Sentinel KQL rules trigger incidents routed to Shuffle SOAR webhook for on-premises SOAR processing |
| DE.AE-03 | Information on adverse events correlated | Multi-source correlation (Splunk + Elastic + Wazuh + network logs); TheHive aggregates alerts from SIEM, EDR, IDS; Shuffle orchestrates cross-platform queries | Wazuh agents on all 8 cloud nodes forward security events to on-premises Splunk and Elastic, enabling unified timeline correlation. Azure Sentinel aggregates Windows Security Events (via AMA DCR home-lab-endpoints) and Syslog (az-ubuntu, gcp-debian-host1) in homelab-log workspace |
| DE.AE-05 | Incident alert thresholds established | Splunk correlation search thresholds; Wazuh rule severity levels; TheHive case severity matrix; Prometheus alert thresholds | AWS CloudWatch Alarms on GuardDuty high/medium findings; GCP Cloud Monitoring alerting policies; Azure Monitor metric alert rules (7 active for az-win2025-dc). Azure Sentinel KQL rule: NSG deny-inbound spike >20 events/hour from single source IP triggers incident |
| DE.AE-07 | Cyber threat intelligence integrated | MISP threat intelligence feeds; CrowdSec community intelligence; Cortex enrichment; Shuffle threat aggregation workflows | AWS GuardDuty ML-based detection integrated with VPC flow, CloudTrail, and DNS query logs. GCP VPC network firewall policy embeds Google Cloud Threat Intelligence (TOR exit nodes, malicious IPs, geo-blocking of sanctioned countries) as priority 100-130 rules. Azure Sentinel threat intelligence connectors feed KQL detection rules |
DE.CM — Continuous Monitoring¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| DE.CM-01 | Networks and network services monitored | DNS query logging; Traefik access logs to SIEM; pfSense flow logs; Suricata/Snort IDS; NetAlertX network monitoring; 100% network traffic visibility | AWS VPC flow logs (fl-03179f74e54bf1aa4) delivered to CloudWatch Logs, capturing accepted and rejected traffic with ENI, source/destination, ports, and action. GCP VPC flow logs enabled on default network (Flow Analyzer). Azure NSG flow logs via Network Watcher (homelab-nsg2). Tailscale tunnel telemetry for all cloud-to-on-premises management paths |
| DE.CM-04 | Malicious code activity monitored | Wazuh FIM; Yara rules; Cortex file analysis; Suricata IDS signatures; ClamAV/Microsoft Defender | Wazuh FIM and rootkit detection deployed on all 8 cloud nodes. Microsoft Defender active on Windows Server 2025 (AWS, Azure). ClamAV on Linux cloud nodes. GuardDuty, GCP SCC, and Azure Defender for Cloud provide behavioral detection independent of signature-based tools |
| DE.CM-06 | External service provider activity monitored | Limited: Cloudflare analytics; VPN logs; public service monitoring; API usage tracking | AWS CloudTrail captures all API calls including IAM role assumptions, security group modifications, and EC2 lifecycle events. GCP Admin Activity logs (always-on) capture all resource changes. Azure Activity Log records all ARM deployments and resource modifications. All cloud control-plane activity available for audit and correlation |
| DE.CM-07 | Monitoring for unauthorized activity performed | Failed authentication tracking; privilege escalation detection; lateral movement monitoring; Shuffle automated analysis | AWS GuardDuty ML detects credential misuse, port scanning, C2 callbacks, and cryptocurrency mining. Azure Sentinel KQL rules detect: NSG deny-inbound spike, Entra ID sign-in from new geographic location, unexpected Windows domain join events, AMA data collection gaps. GCP SCC security health analytics flags open firewall rules and IAM misconfigurations |
| DE.CM-09 | Vulnerability scans performed | Weekly OpenVAS network scans; monthly Nessus authenticated scans; daily PatchMon checks; continuous Wazuh SCA; SIEM correlation of scan results | Wazuh CIS SCA on all 8 cloud nodes daily (CIS Amazon Linux, Debian 12/13, Ubuntu 22.04/24.04, Windows Server 2025). PatchMon daily CVE correlation for 6 Linux cloud nodes. AWS Inspector EC2 agent-based vulnerability assessment. GCP SCC vulnerability assessment for Compute Engine instances. Azure Defender for Cloud recommendations for Azure VMs. All cloud findings visible in unified PatchMon and Wazuh dashboards |
| DE.CM-10 | Threat hunting performed | Wazuh threat hunting queries; Splunk SPL searches; Elastic KQL queries; MITRE ATT&CK-based hunting; Cortex IOC pivoting | CloudWatch Logs Insights queries against CloudTrail and VPC flow logs. GCP Cloud Logging advanced queries for Admin Activity and Data Access anomalies. Azure Sentinel hunting queries (KQL) for cloud-specific TTPs. Cloud Wazuh agent telemetry included in on-premises Splunk SPL correlation searches |
RESPOND (RS) — Tier 3¶
RS.AN — Analysis¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| RS.AN-01 | Notifications investigated | Cortex multi-engine analysis; MISP correlation; Splunk queries; Wazuh forensic data; TheHive case investigation | AWS GuardDuty findings forwarded to CloudWatch Events and SNS — can trigger Lambda-based automated response or Shuffle SOAR webhook. GCP SCC findings generate alerting policy notifications. Azure Sentinel incidents automatically created from KQL rule hits, routed to Shuffle SOAR webhook for on-premises TheHive case creation |
| RS.AN-03 | Forensics performed | Wazuh forensic data collection; memory dumps; network captures; Shuffle automated forensic workflows; evidence preservation | Wazuh agents on cloud nodes collect forensic data (process trees, file hashes, network connections) forwarded to on-premises SIEM. AWS VPC flow logs and CloudTrail provide control-plane forensic trail. GCP Admin Activity logs provide immutable audit trail. Azure Sentinel workspace retains logs for investigation. Cloud node disk snapshots available for offline forensic analysis |
| RS.AN-04 | Incidents categorized | TheHive taxonomy; MITRE ATT&CK mapping; severity scoring; incident classification | AWS GuardDuty finding types mapped to MITRE ATT&CK (e.g., Recon:EC2/Portscan, Impact:EC2/BitcoinDomainRequest, CredentialAccess:IAMUser/AnomalousBehavior). GCP SCC finding categories aligned to cloud attack taxonomy. Azure Sentinel analytics rules tagged with MITRE ATT&CK techniques |
RS.MA — Incident Management¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| RS.MA-03 | Incidents contained | Wazuh Active Response; Cortex responders; Shuffle automated containment workflows; pfSense API firewall rules; network isolation | Wazuh Active Response (firewall-drop, host-deny, account-disable) executes on cloud nodes via Tailscale — same containment playbooks as on-premises. AWS Security Group modifications for EC2 isolation. GCP VPC firewall rule updates via Terraform for rapid containment. Azure NSG rule modification via ARM template or CLI for Azure VM isolation |
| RS.MI-01 | Vulnerabilities mitigated and documented | Virtual patching (SafeLine WAF); IDS signatures; emergency patching via Ansible; TheHive vulnerability tracking | Ansible linux_hardening.yml and update_linux_hosts.yml applied to cloud Linux nodes via Tailscale for emergency patching. GCP Cloud Armor virtual patching capability (WAF rule updates without instance restart). AWS Security Hub findings tracked with remediation status. Azure Defender for Cloud recommendations tracked in homelab-log |
RECOVER (RC) — Tier 3¶
RC.HL — Incident Recovery Plan Execution¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| RC.HL-01 | Recovery plan executed | Backup restoration procedures; system rebuild playbooks; service validation checklists; snapshot rollback | Ansible new_install_baseline_roles.yml provisions replacement cloud nodes from scratch. Tailscale re-enrollment via pre-authorized key. Wazuh agent reinstallation via Ansible playbook. Azure Spot instance restart procedure: start via Azure portal or CLI after Deallocate event. AWS EC2 stop/start preserves Elastic IP association for subnet router continuity |
| RC.HL-02 | Recovery activities completed | Verification testing; service health checks; performance validation; security posture confirmation | Uptime Kuma tunnel health checks confirm Tailscale connectivity restored. Checkmk agent registration confirmed in on-premises dashboard. PatchMon agent check-in verified. Wazuh agent active status confirmed in manager. CIS SCA scan run to confirm security baseline compliance post-rebuild |
RC.RP — Recovery Planning¶
| Subcategory | Description | Implementation | Cloud Extension |
|---|---|---|---|
| RC.RP-01 | Recovery plan exercised | Recovery procedure testing; RTO/RPO validation | Azure Spot instance eviction scenario documented and validated (Deallocate policy, restart procedure). AWS EC2 and GCP Compute instance stop/start tested. Tailscale re-enrollment from pre-authorized key tested. Cloud node RTO target <2hr validated via Ansible rebuild timing. Uptime Kuma alert-to-recovery time documented |
Key Achievements — v2.0¶
- 100% Function Coverage: All six CSF 2.0 functions with documented processes and automated controls, now spanning on-premises and multi-cloud (AWS/Azure/GCP) scope
- Multi-Cloud EDR: Wazuh agents deployed on all 8 cloud nodes with CIS SCA policies matching on-premises baselines; unified vulnerability and compliance dashboards
- Zero Public Management Exposure: No public management ports on any cloud instance — all SSH/RDP access via Tailscale WireGuard with device-bound authentication and ACL enforcement
- Cloud-Native Threat Detection: AWS GuardDuty (ML, VPC flow/CloudTrail/DNS), GCP Security Command Center (asset inventory, vulnerability, network IDS), Azure Defender for Cloud + Sentinel operating in parallel with on-premises SIEM/EDR
- Advanced Detection: Multi-SIEM architecture (Splunk + Elastic + Wazuh) provides <5min MTTD for critical events
- Automated Response: Shuffle orchestration reduces MTTR by 70% compared to manual workflows; Wazuh Active Response containment extended to cloud nodes via Tailscale
- Risk-Based Approach: Same CVSS-based SLAs (Critical <72h, High <7d) enforced across on-premises and cloud scope
- Cloud Recovery Validated: Cloud node rebuild from Ansible playbooks tested; RTO target <2hr achieved; Azure Spot instance eviction/restart procedure documented
Remaining Gaps — Tier 4 Advancement¶
- ML-based anomaly detection / UEBA: No behavioral baseline ML on on-premises nodes; cloud-native ML (GuardDuty, Sentinel) partially fills this gap for cloud nodes
- Formal tabletop exercises and BC/DR drills: No scheduled cross-function exercises; cloud recovery procedures validated informally
- Mutual TLS (mTLS): Not fully deployed for all service-to-service communication, including cloud workloads
- Comprehensive DLP: Egress filtering and bandwidth monitoring in place; formal DLP solution not deployed; cloud egress paths not covered; planned Q2 2026
- Full FIDO2/WebAuthn: TOTP MFA enforced; passwordless WebAuthn not yet deployed
- SBOM tracking: Trivy/Grype implementation planned Q1 2026 for cloud node container and dependency scanning
- Azure Spot eviction automation: Manual restart procedure documented; automated restart workflow not yet implemented
- Cloud UEBA: Azure Sentinel and GCP SCC provide partial behavioral analytics; comprehensive UEBA platform not deployed