📱 App

⚙️ Workers

🗄️ Database

🔌 API

⚡ Cache

🤖 MCP

LAVA®

🪄1-Click Deploy | High Performance N8N setup with Postgres17 + Redis + MCP

N8N 2.0 Pro (Queue Mode with Webhooks, Workers, Code Runners, MCP & Auto-Updates)

worker

Enable Railway private networking between services

Use queue mode for distributed execution processing

🕒 Default timezone used within n8n workflows and nodes

Blocking code nodes from accessing environment variables

This enables AI Agents in n8n to use community node packages as tools

Disable global concurrency limit (use per-worker --concurrency flag)

Sets storage backend for binary data (filesystem, database, or S3)

Offload production webhooks from main process

🔑 Symmetric key for encrypting credentials in database

Path prefix for production webhooks that execute workflows

Path prefix for test webhooks used in the editor

Path prefix for waiting webhooks that resume workflows

📁 Maximum allowed size (in MB) for file uploads via multipart/form-data

⚠️ Allow Git node to use bare repositories (security risk - hooks can execute arbitrary code)

🌐 The domain name where the n8n main instance is accessible

Bind address for incoming connections (IPv6 dual-stack)

📝 Logging verbosity level (set to debug for detailed troubleshooting)

📊 Enable Prometheus metrics endpoint for monitoring n8n instance performance, execution counts, and system health

📈 Include detailed Bull queue statistics (waiting, active, failed) in Prometheus metrics

🐍 Enable the native Python execution environment for Code nodes

📦 Maximum allowed size (in MB) for incoming JSON request payloads

🔒 Force the use of HTTPS for internal URL generation

🔄 Ensure workers immediately pick up workflow changes

🛡️ Shared secret token for authenticating runners with the broker

📍 Address the broker listens on for runner connections

🔌 Port used by the broker to communicate with task runners

Enable task runners for isolated Code node execution

🚀 Maximum number of concurrent tasks a single runner can handle

Maximum payload size in bytes for communication between a task broker and a task runner

🏗️ Enable task runners to process node executions externally

🛡️ Only send session cookies over HTTPS (essential for production)

Skip authentication on OAuth callback endpoints (security risk in n8n 2.0 - defaults to false for enforced auth)

📁 Enable the workflow templates library in the UI

Node.js runtime options (memory limits, GC tuning)

🚪 TCP port where the n8n process listens for HTTP traffic

Enable IPv4/IPv6 dual-stack Redis support

🚦 Enable queue health monitoring and worker status checks

⏳ Grace period (in seconds) to wait for active requests to finish before shutting down old deployment

🌍 System-level timezone for the container environment

Public HTTPS endpoint for incoming webhook triggers

redis

⚠️ Container execution UID (0 for root); ensures permission to write persistence files

Hostname of the private Redis service endpoint inside Railway’s network

🔑 Alias for the Redis password to ensure compatibility with various clients

🚪 TCP port Redis listens on for client connections

👤 Redis username, typically the default ACL user in secured Redis setups

🔐 Randomly generated 64-character strong password for Redis authentication

🌐 External connection string for monitoring Redis via CLI or GUI

Full Redis connection URL for the job queue (Bull)

postgres

Public URL to connect to Postgres database, used by the Data panel.

Full PostgreSQL connection string including credentials, host, port, and database

Filesystem directory where PostgreSQL stores its data files and WAL

Default PostgreSQL database name for client connections

Hostname of the PostgreSQL server on Railway’s private network

🔐 Password for authenticating the PostgreSQL user

🚪 TCP port PostgreSQL listens on for client connections

👤 PostgreSQL user used for client connections

Name of the primary PostgreSQL database created at initialization

🔑 Random 64-character strong password for securing the PostgreSQL user

👤 PostgreSQL username for superuser / application user

⏱️ Grace period in seconds allowing Postgres to cleanly shut down and flush data to disk

Validity period in days for generated SSL certificates (roughly 2.25 years)

🛡️ Links the active authentication token to the generated secret below

🔐 Auto-generated 64-character secret to secure the MCP-to-AI connection

🔇 If true, hides logs to prevent JSON-RPC corruption (keep false unless using Stdio)

📝 Logging verbosity; "error" ensures clean communication between the AI and MCP

🌐 Sets the communication protocol (using HTTP for remote AI access)

🔑 API key from your n8n User Settings to allow MCP to control n8n

🔗 The base URL of your main n8n instance for API calls

🛑 Opt-out of anonymous usage data collection for privacy

🚪 TCP port where the MCP server listens for incoming AI requests

🛡️ "strict" enforces high security; use "moderate" only for local testing

runner

🕒 Timezone inherited for code execution logic

🔒 Security: Prevents code executed in the runner from accessing system environment variables

🧪 Allows modifying JS prototypes (required for certain complex libraries in Code nodes)

🛡️ Shared secret token to authenticate this runner with the Task Broker

⏳ Minutes to wait before shutting down an idle runner (useful for serverless/auto-scaling)

📦 List of allowed external npm/pip packages for use in Code nodes

🚦 Port dedicated to the runner launcher's health check monitoring

📝 Logging verbosity for the runner's launcher process

🚀 Maximum number of code tasks this runner can execute simultaneously

📚 List of allowed Python standard library modules (* allows all)

🔗 Internal URI used to connect the runner to the task broker

🛠️ Built-in Node.js modules allowed in the Code node (e.g., path, fs)

🔌 External npm packages allowed in the Code node

⏳ Grace period (in seconds) to wait for active tasks to finish before shutting down

🌍 System-level timezone for the runner container

main

Database type - using PostgreSQL for data storage

Enable private networking for Railway services communication

Use queue mode with workers for distributed execution

Blocking code-nodes from accessing environment variables

Sets the storage backend for binary data (e.g. files) during executions

Offload production webhook processing to workers instead of main process

🔑 Encryption key for sensitive data stored in database

Path prefix for waiting webhooks that resume waiting workflows when called

Enforce 0600 file permissions on configuration files for security

Listen on all IPv6 addresses (and IPv4 via dual-stack)

📊 Enable Prometheus metrics endpoint for monitoring n8n instance performance

🔒 Protocol used for internal URL generation (ensures generated links use https)

Automatically install missing npm packages for nodes

🔄 Ensure workers immediately pick up changes made in the editor

✅ Enable the use of Task Runners for improved scalability

🏗️ Enable task runners to process node executions outside the main process

Node.js memory limit set to 8GB for handling concurrent executions

Route manual workflow executions (UI "Run" button) to workers instead of main instance

🚪 Port where n8n listens for incoming connections

Enable IPv4 and IPv6 dual-stack support for Redis connections

Public webhook URL for incoming webhook triggers

webhook

Name of the PostgreSQL database n8n will use

Hostname of the PostgreSQL server on the private network

Password for authenticating the PostgreSQL user

TCP port PostgreSQL listens on (usually 5432)

PostgreSQL username used by n8n to connect

Database engine type used by n8n (PostgreSQL in this setup)

Select execution mode; “queue” uses workers via Redis

Allows AI Agents to use community node packages as tools

Disable global concurrency limit (use per-worker flags)

💾 Sets the storage backend for binary data during executions

Offload production webhook processing to workers

🔑 Symmetric key used to encrypt credentials at rest

Enforce strict 0600 permissions on n8n config files

Allow Git node to use bare repositories (security risk)

Bind to all IPv6 interfaces (and IPv4 via dual-stack)

🔒 Protocol used for the instance (set to https for production)

Turn on task runners to isolate Code node execution

🛡️ Only send session cookies over HTTPS (essential for security)

Skip authentication on OAuth callback endpoints

Node.js runtime flags such as memory limits or GC tuning

Send manual “Run” executions to workers

Enable IPv4/IPv6 dual-stack support for Redis

Redis server hostname used for Bull job queues

Redis password for authenticating queue connections

Redis username for ACL‑based authentication

Public HTTPS endpoint used for incoming n8n webhooks