SIEM Docker Stack

⚠️ Active Development — This project is under active development and subject to change. Configuration formats, directory layouts, and integration interfaces may evolve between releases. Pin to a specific commit if you need stability.

A production-ready, fully Dockerized SIEM/SOC stack with hot/warm tiering for home labs and small-to-medium businesses. Designed to run on a single server with two-tier storage (NVMe + SATA) for cost-effective log retention.

Architecture

  ┌─────────────┐                    ┌──────────────────────────────────────────────────────┐
  │   pfSense   │──── Syslog ───────►│              SIEM Docker Stack                       │
  │   Router    │     UDP 514        │                                                      │
  │             │                    │  ┌──────────┐   ┌──────────┐   ┌──────────────────┐  │
  │  Suricata   │── EVE JSON ───────►│  │ Syslog-ng │──►│ Logstash │──►│ OpenSearch (Hot)  │  │
  │  IDS/IPS    │   UDP 5140         │  └──────────┘   └──────────┘   │    NVMe SSD       │  │
  │             │                    │                                │    0-30 days      │  │
  │  CrowdSec   │── Wazuh rules ────►│                                └────────┬───────────┘  │
  │  (edge)     │                    │                                         │ ISM 30d      │
  │             │                    │  ┌──────────┐                   ┌──────▼───────────┐  │
  │  Telegraf   │── InfluxDB ───────►│  │ InfluxDB │                   │ OpenSearch (Warm) │  │
  │  Metrics    │   HTTP 8086        │  └──────────┘                   │    SATA SSD       │  │
  └─────────────┘                    │                                 │    30-365 days    │  │
                                     │  ┌────────────┐                 └──────────────────┘  │
  ┌─────────────┐                    │  │ Prometheus │                                       │
  │   UniFi     │── Poller ─────────►│  └────────────┘                 ┌──────────────────┐  │
  │  Network    │                    │                                 │     Grafana      │  │
  └─────────────┘                    │  ┌──────────────────────────┐   │   Dashboards     │  │
                                     │  │    Wazuh (EDR/SIEM)      │   │   + Alert Rules  │  │
  ┌─────────────┐                    │  │  Manager + Indexer +     │   └────────┬─────────┘  │
  │   Wazuh     │── Agent ──────────►│  │  Dashboard               │            │            │
  │   Agents    │   TCP 1514         │  │  VirusTotal (cached)     │            │            │
  └─────────────┘                    │  │  custom-n8n integration  ├──┐         │            │
                                     │  └──────────────────────────┘  │         │            │
  ┌─────────────┐                    │                                │         │            │
  │  JumpCloud  │── Bridge ─────────►│  ┌──────────────────────────┐  │         │            │
  │  (optional) │   JSONL            │  │       N8N (SOAR)         │◄─┘         │            │
  └─────────────┘                    │  │  Alert Triage (Wazuh)    │◄───────────┘            │
                                     │  │  Alert Router (Grafana)  │                         │
                                     │  │  CrowdSec Enrichment     │──────►  Discord         │
                                     │  └──────────────────────────┘         Notifications   │
                                     │                                                       │
                                     │  ┌──────────────────────────┐                         │
                                     │  │   Portainer (optional)   │                         │
                                     │  └──────────────────────────┘                         │
                                     └───────────────────────────────────────────────────────┘

Features

Hot/Warm Tiering — Automatically migrates indices from NVMe → SATA after 30 days, deletes after 1 year
15+ Services — OpenSearch (2-node hot/warm), Logstash, Grafana, InfluxDB, Prometheus, Wazuh (Manager + Indexer + Dashboard), Syslog-ng, UniFi Poller, N8N, Portainer
N8N SOAR — 3 automation workflows: Wazuh alert triage, Grafana alert router, CrowdSec enrichment — all route to Discord with severity-based filtering and 90s dedup
VirusTotal Caching — SQLite-backed cache in the Wazuh integration layer. Verdict-based TTLs (clean 7d, detected 30d) eliminate redundant API calls for repeated FIM hashes
Pre-built Dashboards — 17 dashboards covering Wazuh security/compliance/agents, SIEM overview, CrowdSec, JumpCloud IdP, Docker, Prometheus, NVIDIA GPU, Twingate ZTNA
CrowdSec Integration — Edge enforcement on pfSense, Wazuh decoding/alerting, Grafana dashboards, n8n enrichment with OpenSearch context lookups
Automated Setup — Numbered scripts (01-06) walk through disk formatting → system tuning → deployment → verification
ISM Lifecycle — Index State Management handles the hot→warm→delete lifecycle automatically
Twingate ZTNA — Zero Trust Network Access connector monitoring with Wazuh rules (120700–120750), access control tracking, STUN/relay health, connector state, and per-connector event breakdown
pfSense Integration — Suricata IDS/IPS logs, pfBlockerNG, syslog, CrowdSec, and Telegraf metrics
JumpCloud IdP — Optional bridge for Directory Insights ingestion with Wazuh decoders/rules
Production-Ready — Docker daemon tuned, kernel parameters optimized, firewall configured, systemd timers for updates

Reference Build

This stack was built and tested on the following hardware. You do not need identical hardware — this is a reference for sizing.

Component	Specification
Server	Supermicro SuperServer 5019A-FTN4 (1U Rackmount)
CPU	Intel Atom C3758 (8 cores, 2.2 GHz, 25W TDP)
RAM	64 GB DDR4 ECC (4 × 16 GB)
OS Disk	240 GB SanDisk SSD PLUS (SATA)
HOT Storage	1 TB Samsung 970 EVO Plus (NVMe)
WARM Storage	2 TB Samsung 870 EVO (SATA SSD)
Network	4 × GbE (Intel I350-AM4)
OS	Ubuntu 24.04 LTS Server

Memory Budget (64 GB Reference)

Service	JVM Heap	Notes
OpenSearch Hot	12 GB	Primary data node, cluster manager
OpenSearch Warm	4 GB	Data-only node for older indices
Wazuh Indexer	4 GB	Separate OpenSearch instance for Wazuh
Logstash	2 GB	Log parsing pipeline
Total JVM	22 GB
OS + Docker + Buffers	~42 GB	File system cache benefits OpenSearch

Minimum Requirements: 16 GB RAM (with reduced heap sizes), 2 CPU cores, 100 GB SSD + any secondary disk. See docs/troubleshooting.md for memory budget guidelines at different RAM levels.

Services

Service	Port	Description
Grafana	3000	Dashboards, visualization, alert rules
OpenSearch (Hot)	9200	Primary log search & indexing (HTTPS + auth, NVMe, 0-30 days)
OpenSearch (Warm)	—	Read-optimized node (SATA, 30-365 days)
OpenSearch Dashboards	5601	OpenSearch UI
Wazuh Manager	1514/udp	Agent enrollment, EDR, FIM, VirusTotal integration
Wazuh Indexer	9200 (internal)	Dedicated OpenSearch for Wazuh alerts
Wazuh Dashboard	443	EDR/SIEM dashboard (HTTPS)
Wazuh API	55000	Wazuh RESTful API
N8N	80	SOAR orchestrator — alert triage, enrichment, Discord routing
Prometheus	9090	Metrics scraping & alerting
Prometheus OTLP	9091 (loopback)	Receiver for pushed OpenTelemetry metrics; authenticated and firewalled, see docs/prometheus-otlp.md
Alertmanager	9093 (loopback)	Independent alert delivery (Prometheus rules → n8n)
elasticsearch-exporter	—	OpenSearch cluster/index metrics for Prometheus (internal :9114)
InfluxDB	8086	Time-series metrics (pfSense, Telegraf, UniFi) — auth enabled
Logstash	5140/udp	Suricata EVE JSON ingestion
Syslog-ng	514/udp+tcp	Centralized syslog receiver
UniFi Poller	—	Collects UniFi switch/AP telemetry → InfluxDB
CrowdSec	—	Optional local LAPI (profile-gated); production runs on pfSense edge
Portainer	9443 (loopback)	Docker management UI. Opt-in: `docker compose --profile portainer up -d`

Note: N8N runs as a separate container (not in the main docker-compose.yml) on a macvlan network. See docs/n8n-soar.md for deployment.

Alert & Notification Architecture

All alerts flow through N8N for dedup, severity filtering, and enrichment before reaching Discord. There is no direct Wazuh→Discord or Grafana→Discord path.

  Wazuh Manager ──(custom-n8n webhook)──► N8N Wazuh Alert Triage
      level 10+                              │
                                             ├─ Critical (12+) ──► Discord (red embed)
                                             ├─ High (10-11) ────► Discord (orange embed)
                                             └─ Medium/Low ──────► Log only (no notification)

  Grafana ──────(contact point webhook)───► N8N Grafana Alert Router
      alert rules                            │
                                             ├─ Firing ──────────► Discord (severity embed)
                                             └─ Resolved ────────► Discord (green embed)

  CrowdSec ────(via Grafana alert rule)───► N8N CrowdSec Enrichment
      ban decisions                          │
                                             ├─ OpenSearch lookup (IP context)
                                             └─ Enriched ────────► Discord (embed with ban details)

Dedup: N8N applies a 90-second burst window per rule ID + source IP to prevent alert storms.

Secrets: Discord webhook credentials are stored in N8N's encrypted credential store, not in config files.

Quick Start

Prerequisites

Ubuntu 22.04+ (or any modern Linux with Docker support)
Two storage devices: one fast (NVMe/SSD for HOT), one large (SATA SSD/HDD for WARM)
At least 16 GB RAM (32+ GB recommended)
Docker CE 24+ with Compose V2

Step-by-Step Installation

# 1. Clone the repository
git clone https://github.com/ChiefGyk3D/siem-docker-stack.git
cd siem-docker-stack

# 2. Format and mount data disks (EDIT the script first!)
#    Change NVME_DEV and SATA_DEV to match your hardware.
#    Run `lsblk` to identify your disks.
sudo nano scripts/01-disk-setup.sh
sudo bash scripts/01-disk-setup.sh

# 3. Install Docker, tune kernel, configure firewall
sudo bash scripts/02-bootstrap.sh

# 4. Configure your environment
cp .env.example .env
nano .env    # Set YOUR IPs, passwords, and heap sizes

# 5. Generate Wazuh TLS certificates
bash scripts/06-generate-wazuh-certs.sh

# 6. Deploy the stack (local mode)
bash scripts/03-deploy.sh local

# 7. Wait ~2 minutes, then apply index templates and ISM policy
bash scripts/04-apply-ism-policy.sh http://localhost:9200

# 8. Verify everything is healthy
bash scripts/05-verify.sh localhost

# 9. Import the Grafana dashboards (file-based provisioning is not wired up —
#    dashboards are deployed via the Grafana API)
python3 scripts/deploy-dashboards.py

After Installation

Web UIs bind to loopback by default (see Network exposure). Reach them from your workstation with an SSH tunnel, for example ssh -L 3000:127.0.0.1:3000 user@siem-host, or widen BIND_ADDR on purpose.

Grafana → http://127.0.0.1:3000 (user admin, password GRAFANA_ADMIN_PASS from .env)
Wazuh Dashboard → https://127.0.0.1:443 (user admin, password WAZUH_INDEXER_PASSWORD from .env)
OpenSearch Dashboards → http://127.0.0.1:5601
Portainer (only if enabled) → https://127.0.0.1:9443

There are no default passwords: compose refuses to start until every required secret in .env is set (see Required secrets).

Changing Default Passwords

An interactive password change script is included to safely update credentials for all stack components:

# Run on the SIEM server
sudo bash /opt/siem/change-passwords.sh

The script handles:

Wazuh Indexer (admin) — Generates bcrypt hash, applies to security config, updates WAZUH_INDEXER_PASSWORD in .env
Wazuh Dashboard (kibanaserver) — Same for the dashboard's service account, updates WAZUH_DASHBOARD_PASSWORD in .env
Wazuh API (wazuh-wui) — Changes via REST API, updates WAZUH_API_PASSWORD in .env and the Dashboard wazuh.yml
Grafana — Resets via Grafana CLI, updates GRAFANA_ADMIN_PASS in .env
Never edits docker-compose.yml; every secret lives in .env (keep it mode 600)
Sends passwords to containers on stdin, not on docker exec command lines (the Grafana CLI reset is the one exception)
Uses docker compose up -d (not restart) so env var changes take effect
Verifies each password change worked before finishing

TIP: Avoid using $ in passwords — it causes escaping issues across YAML, shell, Docker Compose, and JSON layers.

Required secrets

There are no default passwords. Every secret below is required: it ships empty in .env.example, and a missing, empty or placeholder value stops the stack or script that needs it. Values are never printed in an error message; the message names the variable. Use strong, unique values and avoid $.

Before docker compose up, run bash scripts/check-env.sh .env; scripts/03-deploy.sh runs it for you. It rejects an empty value or a placeholder (the old well-known defaults and obvious fill-ins such as your_*, <...>, example*, placeholder*, xxx*, admin and password), compared case-insensitively. The exact rule lives in scripts/lib-secrets.sh, and each scripts/deploy-*.py carries the same one.

Stack secrets (.env, mode 600). Compose interpolates the whole file, so if any of these is unset docker compose refuses to run and names the variable. A value that is set but still a placeholder is caught by check-env.sh, not by compose.

Secret	Protects	Without it
`OPENSEARCH_ADMIN_PASSWORD`	OpenSearch `admin` (main cluster), used by the 04, 05, 08, 09 and 10 scripts	Those scripts exit with an error (05 skips its authenticated checks); `check-env.sh` fails
`OPENSEARCH_LOGSTASH_PASSWORD`	Logstash write account on OpenSearch	Compose refuses to run
`OPENSEARCH_READONLY_PASSWORD`	Read-only OpenSearch account used by Grafana and the Elasticsearch exporter	Compose refuses to run
`OPENSEARCH_DASHBOARDS_PASSWORD`	OpenSearch Dashboards service account	Compose refuses to run
`GRAFANA_ADMIN_PASS`	Grafana `admin` login	Compose refuses to run
`INFLUXDB_ADMIN_PASSWORD`	InfluxDB admin (only InfluxDB itself holds it)	Compose refuses to run
`GRAFANA_INFLUX_PASSWORD`	Read-only InfluxDB account `grafana_ro` for Grafana	Compose refuses to run
`UNPOLLER_INFLUX_PASSWORD`	Write-only InfluxDB account `unpoller_wo` for unifi-poller	Compose refuses to run
`WAZUH_INDEXER_PASSWORD`	Wazuh indexer `admin`, used by the manager and dashboard	Compose refuses to run
`WAZUH_API_PASSWORD`	Wazuh API user `wazuh-wui`	Compose refuses to run
`WAZUH_DASHBOARD_PASSWORD`	Wazuh dashboard service account (`kibanaserver` on the indexer)	Compose refuses to run
`WAZUH_GRAFANA_RO_PASSWORD`	Read-only Wazuh indexer user for Grafana (`scripts/11-wazuh-grafana-readonly.sh` creates it)	Compose refuses to run; script 11 exits with an error
`UNIFI_PASS`	The UniFi controller account unifi-poller logs in with	Compose refuses to run
`PROMETHEUS_OTLP_SENDER_PASSWORD`	Basic-auth account `sender` for OTLP senders on the `prometheus-otlp` receiver	`check-env.sh` fails; the render script exits with an error
`PROMETHEUS_OTLP_GRAFANA_PASSWORD`	Grafana's account `grafana` on `prometheus-otlp`	Compose refuses to run
`PROMETHEUS_GRAFANA_PASSWORD`	Grafana's account `grafana` on the SIEM Prometheus (9090 now requires basic auth)	Compose refuses to run
`PROMETHEUS_SELF_PASSWORD`	Account `self` on the SIEM Prometheus: its own scrape, and operator queries or reload	`check-env.sh` fails; the render script exits with an error
`PROMETHEUS_OTLP_SCRAPE_PASSWORD`	Account `scrape` the main Prometheus uses to scrape `prometheus-otlp`	`check-env.sh` fails; the render script exits with an error

Client-side secrets (environment, not the stack .env). The deploy scripts have no fallback: unset, empty or placeholder, they exit with an error before sending any request.

Secret	Needed by	Without it
`GRAFANA_PASS` (the smoke test also accepts `GRAFANA_ADMIN_PASS`)	`deploy-dashboards.py`, `deploy-grafana-alerts.py`, `deploy-crowdsec-dashboard.py`, `deploy-n8n-soar.sh`, `08-crowdsec-smoketest.sh`	Script exits 1, naming `GRAFANA_PASS`
`N8N_API_KEY`	`deploy-n8n-soar.sh`, `deploy-n8n-grafana-router.py`	Script exits 1
`N8N_PASSWORD`	`deploy-n8n-grafana-router.py` (n8n login)	Script exits 1
`DISCORD_WEBHOOK_URL`	`deploy-n8n-grafana-router.py`	Script exits 1

Where to keep them: .env on the SIEM host for the stack secrets, and your shell, a secrets manager or Doppler for the client-side ones (see Secrets Management). Do not commit either. Optional integration keys (VirusTotal, JumpCloud, alert-delivery webhooks) are not required to run the stack; the feature is simply off until they are set.

Network exposure

Docker-published ports bypass UFW, so the bind address in .env is the real access control. Everything binds to loopback (127.0.0.1) unless you say otherwise:

Variable	Default	Controls
`BIND_ADDR`	`127.0.0.1`	Management and data UIs/APIs: OpenSearch 9200/9201, OpenSearch Dashboards 5601, Grafana 3000, Prometheus 9090, Wazuh API 55000, Wazuh indexer 9202, Wazuh Dashboard 443
`INGEST_BIND_ADDR`	`127.0.0.1`	Log and metric ingestion: syslog 514, Logstash 5140/udp and 5044, Wazuh agents 1514/1515, InfluxDB 8086 (pfSense Telegraf writes)
`PORTAINER_BIND_ADDR`	`127.0.0.1`	Portainer 9443 (only when the `portainer` profile is on)
`PROMETHEUS_OTLP_BIND_ADDR`	`127.0.0.1`	Prometheus OTLP receiver 9091. Authenticated (basic auth) and must be firewalled to approved senders before you widen it: docs/prometheus-otlp.md

To expose on purpose, set the variable to one interface address, not 0.0.0.0 unless you must, then recreate the stack:

# .env -- accept firewall/sensor logs and agents on the LAN interface only
INGEST_BIND_ADDR=192.0.2.10
# .env -- reach the web UIs from the LAN (prefer an SSH tunnel or reverse proxy)
BIND_ADDR=192.0.2.10

scripts/02-bootstrap.sh adds ufw rules only when you pass a source network (SIEM_ALLOWED_CIDR=192.0.2.0/24), and each rule names that source. Because published ports skip UFW, restrict who may reach a widened port with a DOCKER-USER iptables rule or by binding to a management VLAN address.

Portainer (opt-in)

Portainer mounts the Docker socket read-write: control of its UI is root on the host. It does not start by default. To enable it:

cd /opt/siem && docker compose --profile portainer up -d portainer
# then browse https://127.0.0.1:9443 (SSH tunnel), or set PORTAINER_BIND_ADDR

Upgrade notes

This release removes every fallback secret and changes default port binding. Existing deployments must act before docker compose up -d, or compose stops with a message naming the missing variable.

New required variables. Add these to .env (strong, unique values; avoid $). scripts/check-env.sh .env checks for missing or well-known values and scripts/03-deploy.sh runs it for you.

# already existed in .env.example, now REQUIRED (no fallback in compose)
OPENSEARCH_LOGSTASH_PASSWORD=
OPENSEARCH_READONLY_PASSWORD=
OPENSEARCH_DASHBOARDS_PASSWORD=
GRAFANA_ADMIN_PASS=
INFLUXDB_ADMIN_PASSWORD=
WAZUH_INDEXER_PASSWORD=
WAZUH_API_PASSWORD=
UNIFI_PASS=
# new
WAZUH_DASHBOARD_PASSWORD=     # kibanaserver on the Wazuh indexer
WAZUH_GRAFANA_RO_PASSWORD=    # read-only Wazuh indexer user for Grafana
GRAFANA_INFLUX_PASSWORD=      # read-only InfluxDB user for Grafana
UNPOLLER_INFLUX_PASSWORD=     # write-only InfluxDB user for unifi-poller
PROMETHEUS_OTLP_SENDER_PASSWORD=   # OTLP receiver accounts; then run scripts/render-prometheus-web-config.sh
PROMETHEUS_OTLP_GRAFANA_PASSWORD=
PROMETHEUS_OTLP_SCRAPE_PASSWORD=
PROMETHEUS_GRAFANA_PASSWORD=       # basic auth on the SIEM Prometheus (9090)
PROMETHEUS_SELF_PASSWORD=
# optional
BIND_ADDR=127.0.0.1
INGEST_BIND_ADDR=127.0.0.1

OPENSEARCH_ADMIN_PASSWORD is also checked by check-env.sh. If your .env was copied from an old .env.example, replace every well-known default value; check-env.sh rejects them. The full list, with what each secret protects, is in Required secrets.

Log ingestion stops until you set INGEST_BIND_ADDR. Firewalls, sensors and Wazuh agents on other hosts can no longer reach 514, 5140, 5044, 1514, 1515 or 8086 until INGEST_BIND_ADDR is set to the SIEM host's interface address. Likewise the web UIs and APIs are loopback-only until BIND_ADDR is set.
Wazuh dashboard service account. The dashboard no longer uses the built-in kibanaserver password. Put the new value in WAZUH_DASHBOARD_PASSWORD, start the indexer, run sudo bash /opt/siem/change-passwords.sh and enter the same value at the "Wazuh Dashboard (kibanaserver)" prompt, then docker compose up -d. The dashboard cannot reach the indexer between the compose update and that step.
InfluxDB accounts. On the next start the influxdb container creates grafana_ro (read on pfsense, unpoller, telegraf) and unpoller_wo (write on unpoller). Grafana and unifi-poller no longer hold the Influx admin password. Names can be changed with GRAFANA_INFLUX_USER and UNPOLLER_INFLUX_USER. The pfSense Telegraf writer keeps whatever account you gave it; give it its own write-only user rather than the admin.
Portainer no longer starts by default. Enable it with docker compose --profile portainer up -d portainer. A Portainer container from the old stack keeps running until you stop it (docker stop portainer).
change-passwords.sh writes to .env, not into docker-compose.yml. It now requires /opt/siem/.env to exist.
Scripts. 05-verify.sh needs WAZUH_INDEXER_PASSWORD in the environment or .env and no longer expects the portainer container. 02-bootstrap.sh opens no SIEM ports in ufw unless SIEM_ALLOWED_CIDR is set.
Grafana reads the Wazuh indexer with a read-only account. The Wazuh datasource no longer uses the indexer admin user, and Grafana no longer receives WAZUH_INDEXER_PASSWORD. Set WAZUH_GRAFANA_RO_PASSWORD in .env (name configurable with WAZUH_GRAFANA_RO_USER, default grafana_ro), start the stack, then run bash scripts/11-wazuh-grafana-readonly.sh and docker compose up -d grafana. The Wazuh indexer keeps its security config inside the container, so the script bcrypt-hashes the password there (on stdin), adds the user, the wazuh_grafana_readonly role (read on wazuh-alerts-*, -archives-*, -monitoring-*, -statistics-*) and the mapping to the container's internal_users.yml, roles.yml and roles_mapping.yml, applies each with securityadmin.sh, and checks that the account can read and gets 403 on a delete. No hash or password is in a tracked file. Until you run it, Grafana's Wazuh panels fail to log in. Re-run it after recreating the wazuh-indexer container or changing the password. Rotate the password by editing .env, re-running the script and recreating grafana.

No default-stack Grafana datasource uses an admin account any more.

Dashboards

Pre-built Grafana dashboards are included in the dashboards/ directory:

Dashboard	Description
`wazuh_security_overview.json`	Wazuh alerts, agent status, attack distribution
`wazuh_vulnerability_detection.json`	CVE tracking, vulnerable packages, severity breakdown
`wazuh_file_integrity_monitoring.json`	FIM alerts, file changes, affected agents
`wazuh_agent_health.json`	Agent health, SCA compliance, Docker events, VirusTotal, O365 summary
`wazuh_compliance.json`	PCI DSS, NIST 800-53, HIPAA, GDPR compliance dashboards
`wazuh_network_security.json`	MITRE ATT&CK, SSH/Auth, pfSense alerts, GeoIP threat map
`wazuh_office365.json`	Office 365 SharePoint, Exchange, Azure AD, Copilot audit monitoring
`siem_overview.json`	Unified SIEM overview — cross-source correlation, Suricata, Wazuh, pfSense
`siem_plus_overview.json`	Extended SIEM+ overview — adds CrowdSec, JumpCloud, and Office 365 panels
`siem_ztna_overview.json`	Combined SIEM + ZTNA overview — all sources plus Twingate access control, alert distribution, timelines
`twingate_ztna.json`	Twingate ZTNA connector monitoring — access events, STUN health, state transitions, rule breakdown
`crowdsec_overview.json`	CrowdSec ban decisions, bouncer events, LAPI alerts, source analysis
`jumpcloud_security.json`	JumpCloud IdP Security — auth events, service breakdown, user activity, MFA tracking
`docker_container_monitoring.json`	Container CPU, memory, network, disk I/O via cAdvisor/Prometheus
`nvidia_gpu_monitoring.json`	NVIDIA GPU utilization, temperature, memory via DCGM/nvidia_gpu_exporter
`prometheus_stats.json`	Prometheus self-monitoring and scrape targets
`datasources_reference.json`	Quick reference for all configured datasources

Importing Dashboards

File-based dashboard provisioning is not wired up — import via the deploy script:

python3 scripts/deploy-dashboards.py

Or manually via the Grafana API:

# Via Grafana API
for f in dashboards/*.json; do
    curl -X POST "http://admin:${GRAFANA_ADMIN_PASS}@localhost:3000/api/dashboards/db" \
        -H 'Content-Type: application/json' \
        -d "{\"dashboard\": $(cat "$f"), \"overwrite\": true}"
done

Or import via the Grafana UI: Dashboards → New → Import → Upload JSON file.

Hot/Warm Storage Strategy

Day 0         Day 30                Day 365
  │             │                      │
  ▼             ▼                      ▼
┌─────────┐  ┌──────────┐          ┌────────┐
│   HOT   │→ │   WARM   │ ──────→  │ DELETE │
│  NVMe   │  │   SATA   │          │        │
│ (fast)  │  │ (merged) │          │        │
└─────────┘  └──────────┘          └────────┘

HOT (0-30 days): Active writes and queries on NVMe. Zero replicas (single-server setup).
WARM (30-365 days): Force-merged to 1 segment for read-optimized queries on SATA.
DELETE (365+ days): Automatically purged by ISM policy.

See docs/disk-strategy.md for detailed sizing guidelines.

pfSense Integration

This stack is designed as the server-side receiver for a pfSense-based network. For the pfSense-side configuration (Suricata, Telegraf, pfBlockerNG, syslog forwarding), see the companion repository:

pfsense_siem_stack — pfSense packages, Telegraf configuration, SID management, syslog format requirements, and 30+ pages of documentation.

Data Flow Summary

Source	Protocol	Port	Pipeline	Index Pattern
Suricata IDS/IPS	UDP	5140	pfSense → Logstash → OpenSearch	`suricata-*`
pfSense Syslog	UDP	514	pfSense → Syslog-ng → Logstash → OpenSearch	`pfsense-syslog-*`
pfSense Filterlog	UDP	514	pfSense → Syslog-ng → Logstash → OpenSearch	`pfsense-filterlog-*`
pfSense Filterlog	UDP	514	pfSense → Syslog-ng → Wazuh Manager → Wazuh Indexer	`wazuh-alerts-*`
pfBlockerNG	InfluxDB	8086	Telegraf → InfluxDB	`pfblockerng`
pfSense Metrics	InfluxDB	8086	Telegraf → InfluxDB	`pfsense`
UniFi Devices	Poller	—	UniFi Poller → InfluxDB	`unpoller`
Wazuh Agents	TCP	1514	Agent → Wazuh Manager → Wazuh Indexer	`wazuh-*`

Note: pfSense must be configured to send syslog in RFC 5424 format (with RFC 3339 timestamps). Syslog-ng parses RFC 5424 and re-formats as BSD syslog for Wazuh's pre-decoder, which requires the traditional timestamp hostname program[pid]: message format to match the built-in pfSense pf decoder.

CrowdSec (Phase 1)

CrowdSec is supported in this stack, but the recommended production path is pfSense-hosted CrowdSec for edge enforcement.

Primary: install CrowdSec on pfSense (bouncer + log processor, optional LAPI)
This repo: ingest CrowdSec events, alerting, dashboards, and notification workflows
Local crowdsec service profile: optional lab/testing mode only

Optional Local CrowdSec Service (Lab Mode)

cd /opt/siem
docker compose -f docker-compose.yml --profile crowdsec up -d crowdsec

Deploy CrowdSec Alert Rule + Dashboard

DS_WAZUH="<your_wazuh_datasource_uid>" \
ALERT_FOLDER_UID="<your_grafana_folder_uid>" \
GRAFANA_URL="http://localhost:3000" \
GRAFANA_USER="admin" \
GRAFANA_PASS="$GRAFANA_ADMIN_PASS" \
python3 scripts/deploy-grafana-alerts.py

DS_WAZUH="<your_wazuh_datasource_uid>" \
GRAFANA_URL="http://localhost:3000" \
GRAFANA_USER="admin" \
GRAFANA_PASS="$GRAFANA_ADMIN_PASS" \
python3 scripts/deploy-crowdsec-dashboard.py

Run Smoke Test

OPENSEARCH_URL="http://localhost:9200" \
GRAFANA_URL="http://localhost:3000" \
GRAFANA_USER="admin" \
GRAFANA_PASS="$GRAFANA_ADMIN_PASS" \
bash scripts/08-crowdsec-smoketest.sh

The smoke test verifies synthetic CrowdSec events can be indexed and that the CrowdSec Grafana alert rule/dashboard are present.

For pfSense install/config/testing steps, see your pfSense repo runbook:

pfsense_siem_stack/docs/crowdsec-phase1.md

VirusTotal Cache

Wazuh's VirusTotal integration triggers on every FIM (syscheck) alert, sending MD5 hashes to the VT API. With many agents and frequent file changes, this burns through the free-tier rate limit quickly.

The cached virustotal.py in wazuh/integrations/ adds an embedded SQLite layer that caches lookup results with verdict-based TTLs:

Verdict	TTL	Rationale
`clean`	7 days	Known-good files rarely change status
`detected`	30 days	Malicious verdicts are stable
`suspicious`	1 day	Re-check ambiguous results quickly
`unknown`	6 hours	File may not have been scanned yet
`not_found`	1 day	May appear after VT processes it
`error`	10 min	Transient API failures

Deployment: The script lives on the host at /data/hot/wazuh/manager/integrations/virustotal.py and is bind-mounted into the container. This means it survives container upgrades — the host file overlays the container image's stock script. The cache DB is stored at /var/ossec/integrations/cache/vt_cache.db (also on the bind mount).

API Key: Configured in ossec.conf via the standard <integration><api_key> block. Wazuh's integratord passes it to the script as a CLI argument — the key is never hardcoded in the script itself.

Optional JumpCloud Bridge

JumpCloud support is intentionally split into a separate optional repo so non-JumpCloud users are not forced to deploy extra components:

jumpcloud-wazuh-bridge — Polls JumpCloud Directory Insights and writes JSONL for Wazuh ingestion. Includes its own Grafana dashboard, Dockerfile, Doppler support, and CI/CD.

The SIEM+ Overview dashboard in this repo includes summary JumpCloud panels (auth failures, events by service) that pull from the same Wazuh index. The dedicated JumpCloud IdP Security dashboard is also stored in dashboards/jumpcloud_security.json.

Wazuh decoders and rules for JumpCloud events are in wazuh/jumpcloud_decoders.xml and wazuh/jumpcloud_rules.xml.

Secrets Management

Current State

Secrets are stored in multiple locations depending on the component:

Secret	Location	Mechanism
VirusTotal API key	`ossec.conf` on Wazuh Manager	`<integration><api_key>` XML block
Discord webhooks	N8N credential store	Encrypted in N8N's SQLite DB
Grafana admin password	`.env` file	Docker Compose env var
OpenSearch passwords	`.env` + security config	Docker Compose + internal security plugin
Wazuh API password	`.env` + `wazuh.yml`	Docker Compose env var
JumpCloud API key	Doppler (if using bridge)	Doppler CLI / env injection

Doppler Migration (Planned)

The goal is to centralize all secrets in Doppler for rotation, audit logging, and environment-based configs. See docs/roadmap.md — Phase 1B for the full migration plan.

The .env.example includes commented-out DOPPLER_PROJECT and DOPPLER_CONFIG vars. The jumpcloud-wazuh-bridge already supports Doppler-first secret resolution as a reference implementation.

Directory Structure

siem-docker-stack/
├── .env.example                          # Environment template — copy to .env
├── README.md
├── LICENSE
├── docker/
│   ├── docker-compose.yml                # Full stack definition
│   ├── opensearch/
│   │   ├── opensearch-hot.yml            # Hot node config
│   │   ├── opensearch-warm.yml           # Warm node config
│   │   ├── ism-hot-warm-policy.json      # ISM lifecycle policy
│   │   ├── index-template-suricata.json  # Suricata field mappings
│   │   └── index-template-pfblockerng.json
│   ├── logstash/
│   │   ├── logstash.yml                  # Logstash config
│   │   └── pipeline/
│   │       ├── 01-suricata.conf          # Suricata EVE JSON parser
│   │       └── 02-syslog.conf            # Syslog router
│   ├── prometheus/
│   │   └── prometheus.yml                # Scrape targets
│   ├── grafana/
│   │   └── provisioning/
│   │       ├── datasources/
│   │       │   └── datasources.yml       # Auto-provisioned datasources
│   │       └── dashboards/
│   │           └── dashboards.yml        # Dashboard provisioning config
│   ├── syslog-ng/
│   │   └── syslog-ng.conf               # Syslog receiver & router
│   └── wazuh/
│       ├── wazuh-indexer.yml             # Wazuh OpenSearch config
│       ├── opensearch_dashboards.yml     # Wazuh Dashboard config
│       └── certs/                        # TLS certificates (generated)
│           └── README.md
├── scripts/
│   ├── 01-disk-setup.sh                  # Format & mount hot/warm disks
│   ├── 02-bootstrap.sh                   # Install Docker, kernel tuning
│   ├── 03-deploy.sh                      # Deploy stack (local or remote)
│   ├── 04-apply-ism-policy.sh            # Apply ISM policy & templates
│   ├── 05-verify.sh                      # Health check all services
│   ├── 06-generate-wazuh-certs.sh        # Generate Wazuh TLS certs
│   ├── 08-crowdsec-smoketest.sh          # CrowdSec integration smoke test
│   ├── 09-test-crowdsec-pfsense-ingest.sh # Replay pfSense CrowdSec fixture logs
│   ├── deploy-n8n-soar.sh               # Deploy N8N workflows + Grafana alerts
│   ├── deploy-n8n-grafana-router.py      # Deploy Grafana alert router to N8N
│   ├── deploy-grafana-alerts.py          # Deploy Grafana SIEM alert rules
│   ├── deploy-dashboards.py              # Import all dashboards via Grafana API
│   └── deploy-crowdsec-dashboard.py      # Import the CrowdSec dashboard
├── n8n/
│   ├── grafana-alert-router.json         # N8N workflow: Grafana → Discord
│   ├── wazuh-alert-triage.json           # N8N workflow: Wazuh → severity triage → Discord
│   └── crowdsec-alert-enrichment.json    # N8N workflow: CrowdSec → OpenSearch → Discord
├── wazuh/
│   ├── integrations/
│   │   ├── virustotal.py                 # Cached VT integration (SQLite TTL layer)
│   │   └── custom-n8n                    # Wazuh → N8N webhook bridge
│   ├── jumpcloud_decoders.xml            # JumpCloud event decoder
│   ├── jumpcloud_rules.xml               # JumpCloud alert rules
│   └── twingate_rules.xml                # Twingate ZTNA connector rules (120700-120750)
├── change-passwords.sh                   # Interactive password change tool
├── dashboards/
│   ├── siem_overview.json                # Cross-source SIEM correlation
│   ├── wazuh_security_overview.json       # Wazuh alerts & attack distribution
│   ├── wazuh_vulnerability_detection.json # CVE tracking & severity breakdown
│   ├── wazuh_file_integrity_monitoring.json # FIM alerts & file changes
│   ├── wazuh_agent_health.json            # Agent health, SCA, Docker events
│   ├── wazuh_compliance.json              # PCI DSS, NIST, HIPAA, GDPR
│   ├── wazuh_network_security.json        # MITRE, SSH, pfSense, GeoIP map
│   ├── wazuh_office365.json               # O365 audit monitoring & GeoIP map
│   ├── docker_container_monitoring.json   # Container metrics via cAdvisor
│   ├── prometheus_stats.json              # Prometheus self-monitoring
│   ├── siem_plus_overview.json             # Extended SIEM+ with CrowdSec, JumpCloud, O365
│   ├── siem_ztna_overview.json             # Combined SIEM + ZTNA unified overview
│   ├── twingate_ztna.json                  # Twingate ZTNA connector health & access control
│   ├── crowdsec_overview.json              # CrowdSec decisions & bouncer activity
│   ├── jumpcloud_security.json             # JumpCloud IdP Security
│   ├── nvidia_gpu_monitoring.json          # NVIDIA GPU metrics
│   └── datasources_reference.json         # Datasource UID reference
├── docs/
│   ├── roadmap.md                        # Phased development roadmap with status badges
│   ├── detection-ownership.md            # Alert rule → workflow mapping reference
│   ├── alert-enrichment-standard.md      # Enrichment requirements for SOAR
│   ├── n8n-soar.md                       # N8N SOAR deployment & workflow reference
│   ├── disk-strategy.md                  # Hot/warm tiering deep dive
│   ├── maintenance.md                    # Maintenance & backup guide
│   ├── troubleshooting.md               # Common issues & fixes
│   ├── crowdsec.md                       # CrowdSec integration guide
│   ├── jumpcloud.md                      # JumpCloud IdP bridge setup
│   └── gpu-monitoring.md                 # NVIDIA GPU Prometheus exporters
└── media/
  ├── icons/                            # Social media icons for README
  └── screenshots/                      # Stack screenshots used in docs

Screenshots

Grafana Dashboards Home

Wazuh Security Overview

Full Stack Verification Script Output

Twingate ZTNA

SIEM + ZTNA Overview

Roadmap

Documentation

Document	Description
docs/roadmap.md	Phased roadmap with status badges (Wazuh tuning, JumpCloud, CrowdSec, SOAR, LLM, Velociraptor, MISP, Zeek, SOC)
docs/detection-ownership.md	Detection-to-workflow mapping, Grafana alert rules, n8n SOAR reference
docs/alert-enrichment-standard.md	Alert enrichment requirements — context fields, evidence, resolve reasons
docs/n8n-soar.md	N8N SOAR deployment, workflow reference, env var setup
docs/disk-strategy.md	Hot/warm tiering, directory layout, sizing guidelines
docs/maintenance.md	Systemd timers, backup strategy, service commands
docs/troubleshooting.md	Common issues, memory budgets, emergency procedures
docs/crowdsec.md	CrowdSec integration — pfSense deployment, Wazuh rules, n8n enrichment
docs/jumpcloud.md	JumpCloud IdP bridge — setup, Wazuh rules, Doppler secrets
docs/gpu-monitoring.md	NVIDIA GPU monitoring — Prometheus exporters for Linux and Windows
.env.example	All configurable environment variables with descriptions

Related Projects

Repository	Description
pfsense_siem_stack	pfSense-side SIEM integration (Suricata, Telegraf, pfBlockerNG, 30+ docs)
jumpcloud-wazuh-bridge	JumpCloud Directory Insights → Wazuh bridge with Grafana dashboard, Docker, Doppler
PiNodeXMR_Grafana_Dashboard	Monero node monitoring dashboard for Grafana

UniFi Network Monitoring

This stack includes UniFi Poller for collecting UniFi switch and AP telemetry, but UniFi Poller itself is a separate community project — not covered in depth here. I have contributed some dashboard fixes upstream. For setup, configuration, and dedicated UniFi dashboards, see the official project:

UniFi Poller (unpoller) — GitHub — Purpose-built UniFi telemetry collector for Grafana + InfluxDB/Prometheus.

License

This project is licensed under the GNU GPL v2.

Support This Project

If this project is useful to you, consider supporting continued development:

Recurring Support

Patreon

StreamElements Tip

Crypto Tips

	Bitcoin `bc1qztdzcy2wyavj2tsuandu4p0tcklzttvdnzalla`
	Monero `84Y34QubRwQYK2HNviezeH9r6aRcPvgWmKtDkN3EwiuVbp6sNLhm9ffRgs6BA9X1n9jY7wEN16ZEpiEngZbecXseUrW8SeQ`
	Ethereum `0x554f18cfB684889c3A60219BDBE7b050C39335ED`

Author & Socials

_Mastodon

_Bluesky

_Twitch

_YouTube

_Kick

_TikTok

_Discord

_Matrix

Name		Name	Last commit message	Last commit date
.github		.github
dashboards		dashboards
docker		docker
docs		docs
media		media
n8n		n8n
scripts		scripts
tests		tests
wazuh		wazuh
.env.example		.env.example
.gitignore		.gitignore
.gitleaksignore		.gitleaksignore
LICENSE		LICENSE
README.md		README.md
SECURITY.md		SECURITY.md
change-passwords.sh		change-passwords.sh

Latest commit

History

Folders and files

Repository files navigation