Skip to content

Infrastructure Layout

Overview

The SOC SIEM cluster is distributed across 5 physical servers (Proxmox VE). Each server has an SSH alias siem-px1..siem-px5. Each hosts LXC containers (CTs) with dedicated CPU, RAM, and SSD. All servers and containers are in VLAN250 (10.250.0.0/24).


Node Layout

siem-px1 (10.250.0.2) — Wazuh Indexer Primary

Hardware: Intel Xeon Silver 4208, 32 cores, 62 GB RAM, ~426 GB SSD (rpool)

CT ID Service CPU RAM SSD IP Role
701 wazuh-indexer 8c 24 GB 250 GB 10.250.0.10 OpenSearch primary node
Reserve 24c 38 GB ~176 GB

siem-px2 (10.250.0.3) — Manager + Case Management

Hardware: Intel Xeon Silver 4210R, 20 cores, 62 GB RAM, ~412 GB SSD (rpool)

CT ID Service CPU RAM SSD IP Role
702 wazuh-manager 6c 8 GB 50 GB 10.250.0.11 Log collection, rules, correlation, integrations
710 thehive 2c 8 GB 30 GB 10.250.0.30 Case management (TheHive 4.1.24)
711 cortex 2c 4 GB 20 GB 10.250.0.31 Automated IoC analysis (Cortex 4.0.1)
Reserve 10c 42 GB ~312 GB

siem-px3 (10.250.0.4) — Dashboard + Indexer Replica 1

Hardware: Intel Xeon E-2334, 8 cores, 62 GB RAM, ~424 GB SSD (rpool)

CT ID Service CPU RAM SSD IP Role
703 wazuh-dashboard 1c 2 GB 10 GB 10.250.0.12 Wazuh web UI (stateless)
704 wazuh-indexer-r1 4c 24 GB 250 GB 10.250.0.13 OpenSearch replica 1
Reserve 3c 36 GB ~164 GB

siem-px4 (10.250.0.5) — Threat Intel + SOAR

Hardware: Intel Xeon E-2334, 8 cores, 62 GB RAM, ~417 GB SSD (rpool)

CT ID Service CPU RAM SSD IP Role
706 misp 2c 4 GB 50 GB 10.250.0.15 Threat Intelligence (MISP 2.5.36)
707 crowdsec 1c 1 GB 10 GB 10.250.0.16 CrowdSec LAPI + blocklist-mirror
712 shuffle 1c 4 GB 20 GB 10.250.0.32 SOAR (Shuffle, currently idle)
Reserve 4c 53 GB ~337 GB

siem-px5 (10.250.0.6) — Indexer Replica 2 + Forensics

Hardware: Intel Xeon E-2334, 8 cores, 62 GB RAM, ~425 GB SSD (rpool)

CT ID Service CPU RAM SSD IP Role
705 wazuh-indexer-r2 4c 24 GB 250 GB 10.250.0.14 OpenSearch replica 2
713 velociraptor 2c 4 GB 50 GB 10.250.0.33 DFIR (Velociraptor 0.76.2)
Reserve 2c 34 GB ~125 GB

Resource Summary

Resource Allocated to CTs Total (5 nodes) Reserve % headroom
CPU cores 33c 76c 43c 57%
RAM 107 GB 310 GB 203 GB 65%
SSD (CT rootfs) 960 GB ~2.1 TB ~1.1 TB 54%
Containers 11

Wazuh Indexer — 3-node cluster (OpenSearch)

Node CT ID Host Role SSD RAM
wazuh-indexer (primary) 701 siem-px1 Master + Data 250 GB 24 GB
wazuh-indexer-r1 (replica) 704 siem-px3 Data 250 GB 24 GB
wazuh-indexer-r2 (replica) 705 siem-px5 Data 250 GB 24 GB

Total cluster capacity: 750 GB SSD (~375 GB usable with 1 replica).

The 3-node cluster provides:

  • Fault tolerance — any single node can fail without data loss
  • Read scaling — queries can be served from any replica
  • Split-brain protection — 3 nodes = proper quorum

Disk Layout (all nodes)

All 5 nodes use ZFS:

Pool Disks Purpose
rpool SSD (stripe or mirror depending on node) Root FS, container rootfs, Docker volumes

HDD datapool (warm tier for log rotation) — not yet created. Planned under TASK-092f (retention policy).


Additional Infrastructure

Service IP Purpose
siem-qnap (QNAP NAS4CD21A) 10.250.0.7 SOC backups: NFS mount /mnt/qnap-soc on all 5 nodes. 2×2TB HDD RAID0, 3.1 TB.

NFS mount is persistent via fstab (_netdev). Proxmox storage qnap-soc-backups visible on all cluster nodes for vzdump.


Last updated: 2026-04-16.