Files

Hutson 38a7a2c52e Auto-sync: 20260123-015626

2026-01-23 01:56:27 -05:00

8.2 KiB

Raw Permalink Blame History

n8n - Workflow Automation

n8n is an extendable workflow automation tool deployed on docker-host2 for automating tasks across your homelab and external services.

Quick Reference

Setting	Value
URL	https://n8n.htsn.io
Local IP	10.10.10.207:5678
Server	docker-host2 (PVE2 VMID 302)
Database	PostgreSQL (containerized)
API Endpoint	http://10.10.10.207:5678/api/v1/

Claude Code Integration (MCP)

n8n-MCP Server

The n8n-MCP server gives Claude Code deep knowledge of all 545+ n8n nodes, enabling it to build complete workflows from natural language descriptions.

Installation: Already configured in ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "n8n-nodes": {
      "command": "npx",
      "args": ["-y", "@czlonkowski/n8n-mcp"]
    }
  }
}

What This Enables:

✅ Build n8n workflows from natural language
✅ Get detailed help with node parameters and options
✅ Best practices for n8n node usage
✅ Debug workflow issues with full node context

Example Prompts:

"Create an n8n workflow to monitor Prometheus and send Telegram alerts"
"Build a workflow that triggers when Syncthing has errors"
"What's the best n8n node to parse JSON responses?"

How It Works:

MCP server provides offline documentation for all n8n nodes
No connection to your n8n instance required
Claude builds workflows that you can then import into https://n8n.htsn.io

Resources:

API Access

API Key

X-N8N-API-KEY: eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiI3NTdiMDA5YS1hMjM2LTQ5MzUtODkwNS0xZDY1MjYzZWE2OWYiLCJpc3MiOiJuOG4iLCJhdWQiOiJwdWJsaWMtYXBpIiwiaWF0IjoxNzY2ODEwMzA3fQ.RIZAbpDa7LiUPWk48qOscJ9-d9gRAA0afMDX_V3oSVo

API Examples

List Workflows:

curl -H "X-N8N-API-KEY: eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiI3NTdiMDA5YS1hMjM2LTQ5MzUtODkwNS0xZDY1MjYzZWE2OWYiLCJpc3MiOiJuOG4iLCJhdWQiOiJwdWJsaWMtYXBpIiwiaWF0IjoxNzY2ODEwMzA3fQ.RIZAbpDa7LiUPWk48qOscJ9-d9gRAA0afMDX_V3oSVo" \
  http://10.10.10.207:5678/api/v1/workflows

Get Workflow by ID:

curl -H "X-N8N-API-KEY: YOUR_API_KEY" \
  http://10.10.10.207:5678/api/v1/workflows/{id}

Trigger Workflow:

curl -X POST \
  -H "X-N8N-API-KEY: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"data": {"key": "value"}}' \
  http://10.10.10.207:5678/api/v1/workflows/{id}/execute

API Documentation: https://docs.n8n.io/api/

Deployment Details

Docker Compose

Location: /opt/n8n/docker-compose.yml on docker-host2

Services:

n8n - Main application (port 5678)
postgres - Database backend

Volumes:

n8n_data - Workflow data, credentials, settings
postgres_data - Database storage

Environment Configuration

N8N_HOST: n8n.htsn.io
N8N_PORT: 5678
N8N_PROTOCOL: https
NODE_ENV: production
WEBHOOK_URL: https://n8n.htsn.io/
GENERIC_TIMEZONE: America/Los_Angeles
DB_TYPE: postgresdb
DB_POSTGRESDB_HOST: postgres
DB_POSTGRESDB_DATABASE: n8n
DB_POSTGRESDB_USER: n8n
DB_POSTGRESDB_PASSWORD: n8n_secure_password_2024

Resource Limits

Memory: 512MB-1GB (soft/hard)
CPU: Shared (4 vCPUs on host)

Common Tasks

Restart n8n

ssh docker-host2 'cd /opt/n8n && docker compose restart n8n'

View Logs

ssh docker-host2 'docker logs -f n8n'

Backup Workflows

Workflows are stored in PostgreSQL. To backup:

ssh docker-host2 'docker exec n8n-postgres pg_dump -U n8n n8n > /tmp/n8n-backup-$(date +%Y%m%d).sql'

Update n8n

ssh docker-host2 'cd /opt/n8n && docker compose pull n8n && docker compose up -d n8n'

Traefik Configuration

File: /etc/traefik/conf.d/n8n.yaml on CT 202

http:
  routers:
    n8n-secure:
      entryPoints:
        - websecure
      rule: "Host(`n8n.htsn.io`)"
      service: n8n
      tls:
        certResolver: cloudflare
      priority: 50

    n8n-redirect:
      entryPoints:
        - web
      rule: "Host(`n8n.htsn.io`)"
      middlewares:
        - n8n-https-redirect
      service: n8n
      priority: 50

  services:
    n8n:
      loadBalancer:
        servers:
          - url: "http://10.10.10.207:5678"

  middlewares:
    n8n-https-redirect:
      redirectScheme:
        scheme: https
        permanent: true

Monitoring

Prometheus

n8n exposes metrics at http://10.10.10.207:5678/metrics (if enabled)

Grafana

n8n metrics can be visualized in Grafana dashboards

Uptime Monitoring

Add to Pulse: https://pulse.htsn.io

Monitor: https://n8n.htsn.io
Check interval: 60s

Troubleshooting

n8n won't start

ssh docker-host2 'docker logs n8n | tail -50'
ssh docker-host2 'docker logs n8n-postgres | tail -50'

Database connection issues

# Check postgres health
ssh docker-host2 'docker exec n8n-postgres pg_isready -U n8n'

# Restart postgres
ssh docker-host2 'cd /opt/n8n && docker compose restart postgres'

SSL/HTTPS issues

# Check Traefik config
ssh root@10.10.10.250 'cat /etc/traefik/conf.d/n8n.yaml'

# Reload Traefik
ssh root@10.10.10.250 'systemctl reload traefik'

API not responding

# Test API locally
curl -H "X-N8N-API-KEY: YOUR_KEY" http://10.10.10.207:5678/api/v1/workflows

# Check if n8n container is healthy
ssh docker-host2 'docker ps | grep n8n'

Remove "This message was sent automatically by n8n" signature from Telegram messages

Problem: n8n Telegram node adds attribution signature to all messages by default.

Solution: Use the correct parameter name appendAttribution (camelCase, not snake_case) in additionalFields:

# Get workflow
curl -H "X-N8N-API-KEY: $(cat /tmp/n8n-key.txt)" \
  http://10.10.10.207:5678/api/v1/workflows/WORKFLOW_ID > workflow.json

# Update all Telegram nodes (using jq)
cat workflow.json | jq '.nodes = (.nodes | map(
  if .type == "n8n-nodes-base.telegram" then
    .parameters.additionalFields.appendAttribution = false
  else
    .
  end
))' | jq '{name, nodes, connections, settings, staticData}' > workflow-fixed.json

# Upload updated workflow
curl -X PUT \
  -H "X-N8N-API-KEY: $(cat /tmp/n8n-key.txt)" \
  -H 'Content-Type: application/json' \
  -d @workflow-fixed.json \
  http://10.10.10.207:5678/api/v1/workflows/WORKFLOW_ID

# Restart n8n to reload workflow
ssh docker-host2 'cd /opt/n8n && docker compose restart n8n'

Important Notes:

Parameter must be appendAttribution (camelCase), not append_attribution or append_n8n_attribution
Must restart n8n after updating workflow for changes to take effect
This applies to all Telegram message nodes in the workflow

Fixed: 2026-01-23

Integration Examples

Homelab Automation Ideas

Backup Notifications - Send Telegram alerts when backups complete
Server Monitoring - Query Prometheus and alert on high CPU/memory
Media Management - Trigger Sonarr/Radarr downloads
Home Assistant Integration - Automate smart home workflows
Git Webhooks - Deploy changes from Gitea automatically
Syncthing Monitoring - Alert when sync folders get behind
UPS Alerts - Notify on power events from NUT

Security Notes

API key provides full access to all workflows and data
Store API key securely (added to this doc for homelab reference)
n8n credentials are encrypted at rest in PostgreSQL
HTTPS enforced via Traefik
No public internet exposure (only via Tailscale)

Quick Start

New to n8n? Start here: N8N-INTEGRATIONS.md ⭐

This guide includes:

✅ Network access verification
✅ Credential setup for all homelab services
✅ 10 ready-to-use starter workflows
✅ Home Assistant, Prometheus, Syncthing, Telegram integrations
✅ Troubleshooting tips

n8n Homelab Integrations Guide - START HERE
docker-host2 VM details
Traefik reverse proxy
IP Assignments
Pulse Setup

Last Updated: 2025-12-26

8.2 KiB Raw Permalink Blame History