Chapter 14: n8n Cloud vs. Self-Hosted n8n (Docker/AWS EC2)¶

Video: Watch this chapter on YouTube (2:54:14)

Overview¶

This chapter explores deployment options for n8n: the managed n8n Cloud service versus self-hosting using Docker on local machines or AWS EC2 instances. Each approach has distinct advantages and tradeoffs.

Detailed Summary¶

The Deployment Decision¶

When starting with n8n, a key decision is how to deploy it:

n8n Cloud: Managed service from n8n
Self-Hosted: Run n8n on your own infrastructure

n8n Cloud: Managed Service¶

Advantages¶

Benefit	Description
Simplicity	Sign up and start building immediately
No infrastructure	No servers, Docker, or DevOps required
Automatic updates	Always on latest version
Security patches	Handled automatically
Official support	Access to n8n support team
Uptime monitoring	Built-in reliability tracking

Disadvantages¶

Drawback	Description
Cost	Monthly subscription fees
Limited control	Can't customize infrastructure
Resource limits	Bound to plan tiers
Data location	May not meet compliance needs

Self-Hosted: Full Control¶

Advantages¶

Benefit	Description
Total freedom	Control resources, location, access
Cost efficiency	Can be cheaper with existing infrastructure
Internal integration	Connect to internal tools/databases
Security compliance	Keep data on-premises
No vendor lock-in	Full ownership of deployment

Disadvantages¶

Drawback	Description
Maintenance burden	You handle updates, patches
No safety net	You fix what breaks
DevOps required	Need infrastructure skills
Uptime responsibility	You ensure availability

Self-Hosting Option 1: Docker Desktop (Local)¶

Prerequisites¶

Docker Desktop: Download from docker.com
Git: For cloning the repository

Installation Steps¶

Clone the repository:

git clone https://github.com/n8n-io/self-hosted-ai-starter-kit.git
cd self-hosted-ai-starter-kit

Set up environment:
```
cp .env.example .env
```
Run Docker Compose (for Mac/CPU):
```
docker compose --profile cpu up
```

The package includes: - n8n - OLLAMA (local LLM) - PostgreSQL (database) - Qdrant (vector database)

Accessing Local n8n¶

Open browser: localhost:5678
Create account (local only, not saved permanently)
Access dashboard

Demo Workflow¶

The starter kit includes a demo workflow: - Chat trigger connected to OLLAMA - Demonstrates local LLM integration - Add fallback model for error handling

Docker Desktop Management¶

View running containers
Start/stop containers
Check logs
Configure resources
Access container terminals

Self-Hosting Option 2: AWS EC2¶

Prerequisites¶

AWS account
Basic EC2 knowledge
SSH client

Using CodeCloud Playground¶

Go to codecloud.com/playground
Select AWS Sandbox
Launch playground
Copy credentials

Creating EC2 Instance¶

Navigate to EC2 in AWS Console
Launch Instance:
Name: "n8n-demo"
AMI: Ubuntu
Instance type: t2.medium (for OLLAMA)
Storage: 30GB
Create Key Pair:
Name: "n8n-demo-key"
Format: .pem
Security Group:
Allow SSH (port 22)
Allow Custom TCP (port 5678) from anywhere
Launch

SSH Into Instance¶

chmod 400 n8n-demo-key.pem
ssh -i n8n-demo-key.pem ubuntu@[PUBLIC_IP]

Install Docker¶

sudo apt update
sudo apt install -y docker.io
sudo systemctl enable docker
sudo usermod -aG docker ubuntu

Install Docker Compose¶

sudo apt install docker-compose-v2
docker compose version

Clone and Configure¶

git clone https://github.com/n8n-io/self-hosted-ai-starter-kit.git
cd self-hosted-ai-starter-kit
cp .env.example .env

Security Configuration¶

Edit .env file to disable secure cookies (for HTTP access):

nano .env

Add:

N8N_SECURE_COOKIE=false

Launch n8n¶

docker compose --profile cpu up -d

Verify Containers¶

docker ps

Should show: n8n, OLLAMA, PostgreSQL, Qdrant

Access n8n¶

Get EC2 public IP from AWS console
Open browser: http://[PUBLIC_IP]:5678
Create account
Start building workflows

Comparison Summary¶

Factor	n8n Cloud	Self-Hosted
Setup time	Minutes	Hours
Maintenance	None	Ongoing
Cost	Subscription	Infrastructure
Control	Limited	Full
Scaling	Plan-based	Manual
Security	Managed	Your responsibility
Updates	Automatic	Manual
Support	Official	Community

Recommendation by Use Case¶

Choose n8n Cloud if: - Just learning n8n - Don't want infrastructure burden - Need reliable uptime - Have budget for subscription

Choose Self-Hosted if: - Need data on-premises - Have DevOps resources - Want full customization - Cost-conscious with existing servers - Enterprise security requirements

Key Takeaways¶

No wrong answer: Both approaches are valid—choose based on needs.
Cloud is easiest start: Sign up and build without infrastructure concerns.
Self-hosting offers control: Full access to environment and data.
Docker simplifies self-hosting: Container-based deployment is straightforward.
Starter kit includes AI tools: OLLAMA, PostgreSQL, Qdrant bundled together.
EC2 enables cloud self-hosting: AWS infrastructure with self-managed n8n.
Security groups are critical: Must allow port 5678 for n8n access.
t2.medium minimum for OLLAMA: Local LLM needs adequate resources.
Maintenance is ongoing: Self-hosting means continuous responsibility.
Hybrid possible: Some organizations use both for different purposes.

Conclusion¶

The deployment choice between n8n Cloud and self-hosting reflects broader tradeoffs in software operations. n8n Cloud removes operational burden at the cost of monthly fees and some flexibility. Self-hosting provides complete control and potential cost savings but requires DevOps expertise and ongoing maintenance. For learning and prototyping, n8n Cloud's simplicity is valuable. For production workloads with specific compliance or integration requirements, self-hosting may be necessary. Many organizations start with Cloud and migrate to self-hosted as needs evolve, or maintain both for different use cases. The detailed Docker and EC2 setup instructions enable learners to experience self-hosting firsthand and make informed decisions.