MCP Issue connect with the VAPI portal

Questions
1

Hi everyone,

image
image491×841 34.1 KB

I’m having trouble getting the Model Context Protocol (MCP) server to work on n8n version 2.3.5 (Self-hosted Docker).

My Setup:

  • n8n Version: 2.3.5

  • Deployment: Docker Compose in Queue Mode (1 Main + 4 Workers).

  • Infrastructure: AWS EC2 behind an Nginx Reverse Proxy.

  • MCP Node: Using the MCP Server Trigger.

The Issue:
Even with N8N_FEATURE_FLAG_MCP=true enabled on both Main and Workers, and N8N_TRUST_PROXY=true set, I am seeing the following error in my logs:
ValidationError: The 'X-Forwarded-For' header is set but the Express 'trust proxy' setting is false (default). Code: ERR_ERL_UNEXPECTED_X_FORWARDED_FOR

This seems to be preventing the MCP SSE stream from establishing.

Nginx Config Snippet:
I have already disabled buffering and gzip for the /mcp/ path:

nginx

server {
    server_name example.com;

    # === MCP Location Block ===
    location /mcp/ {
        proxy_pass http://127.0.0.1:5678;

        # CRITICAL: Disable compression for SSE
        gzip off; 

        proxy_http_version 1.1;
        proxy_set_header Connection ""; 
        proxy_buffering off;           
        proxy_cache off;
        chunked_transfer_encoding on;
        proxy_read_timeout 3600s;      

        # Required for n8n to recognize the proxy protocol
        proxy_set_header Host $host;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
        
        # Prevent Nginx from timing out the SSE stream early
        proxy_connect_timeout 10s;
        proxy_send_timeout 3600s;
    }

    # === Normal n8n UI & API ===
    location / {
        proxy_pass http://127.0.0.1:5678;

        # Ensure headers are correctly forwarded
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
        proxy_set_header X-Forwarded-Host $host;

        # Optional: Handle WebSocket connections (if applicable)
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";

        # Timeouts to avoid 504 errors
        proxy_connect_timeout 3600s;
        proxy_send_timeout 3600s;
        proxy_read_timeout 3600s;

    }
2

The issue appears to be related to the MCP server not starting correctly, despite the feature flag and proxy settings. You should double-check your Docker Compose file to ensure the environment variables `N8N_FEATURE_FLAG_MCP` and `N8N_TRUST_PROXY` are correctly set for both the main n8n instance and all worker instances. Also, confirm your Nginx configuration correctly forwards traffic to the n8n application.

3

Ok, thank you for your guidance. I have already configured all the items you mentioned, as you can see in the docker-compose file and the Nginx virtual host configuration below.

nginx

server {
server_name example.com;

# === MCP Location Block ===
location /mcp/ {
    proxy_pass http://127.0.0.1:5678;

    # CRITICAL: Disable compression for SSE
    gzip off; 

    proxy_http_version 1.1;
    proxy_set_header Connection ""; 
    proxy_buffering off;           
    proxy_cache off;
    chunked_transfer_encoding on;
    proxy_read_timeout 3600s;      

    # Required for n8n to recognize the proxy protocol
    proxy_set_header Host $host;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Proto $scheme;
    
    # Prevent Nginx from timing out the SSE stream early
    proxy_connect_timeout 10s;
    proxy_send_timeout 3600s;
}

# === Normal n8n UI & API ===
location / {
    proxy_pass http://127.0.0.1:5678;

    # Ensure headers are correctly forwarded
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Proto $scheme;
    proxy_set_header X-Forwarded-Host $host;

    # Optional: Handle WebSocket connections (if applicable)
    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";

    # Timeouts to avoid 504 errors
    proxy_connect_timeout 3600s;
    proxy_send_timeout 3600s;
    proxy_read_timeout 3600s;

}

docker-compose file

version: “3.8”

services:
n8n:
image: Package n8n · GitHub
container_name: n8n
restart: unless-stopped
ports:
- “5678:5678”
environment:
- NODE_ENV=production
- N8N_INSTANCE_NAME=prod-n8n
- GENERIC_TIMEZONE=UTC
- N8N_HOST=example.com
- N8N_PORT=5678
- N8N_PROTOCOL=https
- WEBHOOK_URL=https://example.com
- N8N_TRUST_PROXY=true
- N8N_PROXY_HOPS=1
- N8N_SECURE_COOKIE=true
- EXECUTIONS_MODE=queue
- OFFLOAD_MANUAL_EXECUTIONS_TO_WORKERS=true

  # === MCP FIXES FOR MAIN ===
  - N8N_FEATURE_FLAG_MCP=true
  - N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true
  # ==========================

  - QUEUE_BULL_REDIS_HOST=redis
  - QUEUE_BULL_REDIS_PORT=6379
  - DB_TYPE=postgresdb
  - DB_POSTGRESDB_HOST=postgres
  - DB_POSTGRESDB_PORT=5432
  - DB_POSTGRESDB_DATABASE=n8n
  - DB_POSTGRESDB_USER=n8n_postgres
  - DB_POSTGRESDB_PASSWORD=${POSTGRES_PASSWORD}
  - N8N_ENCRYPTION_KEY=${N8N_ENCRYPTION_KEY}
  - N8N_BLOCK_ENV_ACCESS_IN_NODE=true # Kept TRUE for security as requested
  - N8N_RUNNERS_ENABLED=false
  - N8N_REINSTALL_MISSING_PACKAGES=true
  - N8N_NATIVE_PYTHON_RUNNER=false
volumes:
  - n8n_data:/home/node/.n8n
depends_on:
  - postgres
  - redis

n8n-worker:
image: Package n8n · GitHub
command: worker
restart: unless-stopped
environment:
- NODE_ENV=production
- EXECUTIONS_MODE=queue

  # === MCP FIXES FOR WORKERS ===
  - N8N_FEATURE_FLAG_MCP=true
  - N8N_REINSTALL_MISSING_PACKAGES=true
  - N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true
  # =============================

  - QUEUE_BULL_REDIS_HOST=redis
  - QUEUE_BULL_REDIS_PORT=6379
  - DB_TYPE=postgresdb
  - DB_POSTGRESDB_HOST=postgres
  - DB_POSTGRESDB_PORT=5432
  - DB_POSTGRESDB_DATABASE=n8n
  - DB_POSTGRESDB_USER=n8n_postgres
  - DB_POSTGRESDB_PASSWORD=${POSTGRES_PASSWORD}
  - N8N_ENCRYPTION_KEY=${N8N_ENCRYPTION_KEY}
depends_on:
  - redis
  - postgres
deploy:
  replicas: 4

redis:
image: redis:7.2
container_name: redis
restart: unless-stopped
command: redis-server --save “” --appendonly no

postgres:
image: postgres:15
container_name: postgres
restart: unless-stopped
environment:
- POSTGRES_DB=n8n
- POSTGRES_USER=n8n_postgres
- POSTGRES_PASSWORD=${POSTGRES_PASSWORD}
volumes:
- postgres_data:/var/lib/postgresql/data

volumes:
n8n_data:
postgres_data:

1 Like
4

Can you restart your instance now and let me know if anything changes?