Muhammad Ismail Khalid

n8n System Requirements: What You Need to Self-Host n8n

By , Full Stack Developer and AI Automation Engineer. . 6 min read.

Topic: n8n and Automation

On this page
  1. n8n system requirements at a glance
  2. How much RAM does n8n need?
  3. How much CPU does n8n need?
  4. Storage and database
  5. Software requirements
  6. When you need queue mode
  7. Recommended server sizes by workload
  8. Self-hosted n8n or n8n Cloud?
  9. The short answer
  10. Frequently asked questions

n8n has no fixed minimum spec. It will start on very small hardware, and it will also run out of memory on a large server if a workflow pulls in too much data at once. The right size depends on what your workflows do.

This guide covers what n8n's documentation actually says, what drives resource use, and the server sizes I'd start with for different workloads. I build and run n8n workflows for my agency and for clients, including an outreach system that runs on a schedule every day.

n8n system requirements at a glance

n8n system requirements at a glance: RAM, CPU, storage, database, software and network
n8n system requirements at a glance.
Resource What n8n needs A sensible starting point
Memory (RAM) 320 MB to 2 GB in n8n's illustrative sizing 2 GB for light use, 4 GB or more for files and large data
CPU Rarely the bottleneck 1 to 2 vCPUs
Storage 512 MB to 4 GB of SSD for the database in n8n's illustrative sizing 20 GB or more of SSD so logs and execution data have room
Database SQLite by default, PostgreSQL supported SQLite for a single small instance, PostgreSQL for production
Software Docker (recommended install method) Docker Engine and Docker Compose on Linux
Network Port 5678 by default HTTPS through a reverse proxy

The figures in the middle column come from n8n's own sizing guidance, which it bases on n8n Cloud and describes as illustrative only. The right-hand column is my recommendation for a self-hosted server that clients will depend on.

How much RAM does n8n need?

Memory is the resource that matters most. n8n's documentation says memory usually matters more than CPU for planning, and that an idle n8n Cloud instance uses about 100 MB.

What pushes memory up is the work your workflows do. n8n's documentation on memory errors lists these causes:

  • The amount of JSON data a workflow handles
  • The size of any binary data, such as files and images
  • The number of nodes in a workflow
  • Memory-heavy nodes, especially the Code node
  • Manual executions, because n8n copies the data for the editor
  • Other workflows running at the same time

n8n doesn't limit how much data a node can fetch. A workflow that pulls 10,000 rows from a CRM in one step can crash an instance that handles 200 rows without trouble.

Signs n8n is running out of memory

The usual error in the server logs is Allocation failed - JavaScript heap out of memory. You may also see executions that stop without finishing, or the editor becoming slow or unresponsive.

How to reduce n8n memory use

Before you pay for a bigger server, try the fixes n8n recommends:

  1. Process data in batches. Handle 200 rows per run instead of 10,000.
  2. Use sub-workflows. Pair Loop Over Items with Execute Workflow, and have each sub-workflow return only a small amount of data. Memory from each batch is freed when it finishes.
  3. Avoid the Code node where possible. Built-in nodes usually use less memory.
  4. Don't test large jobs manually. Run them in production mode instead.
  5. Raise the Node.js memory limit. Set --max-old-space-size through the NODE_OPTIONS environment variable, but only if the server actually has that memory free.

How much CPU does n8n need?

CPU is rarely the limit. Most workflows spend their time waiting on APIs, not computing. One or two vCPUs handle typical lead routing, CRM updates and notifications well.

CPU matters more when you run many workflows at the same time, transform large files or run heavy Code node logic. If CPU stays high for long periods, that's the point to look at queue mode and workers.

Storage and database

SQLite (the default)

If you don't configure a database, n8n uses SQLite and stores it at ~/.n8n/database.sqlite. It needs no setup and works well for a single small instance.

PostgreSQL

n8n supports PostgreSQL. Its documentation lists support for the two latest actively maintained major versions, plus version 16. Other PostgreSQL-compatible databases, such as CockroachDB and YugabyteDB, aren't supported.

I recommend PostgreSQL for any instance that runs client workflows. It's easier to back up while running, handles more data, and is required if you move to queue mode later.

Plan for execution data

Every execution can be saved with its data. Over months, this is what fills the disk, not n8n itself. Turn on execution data pruning, and decide how long you need to keep successful runs.

Software requirements

n8n recommends Docker. The official Docker setup:

  • Maps n8n to port 5678
  • Stores your data in a volume mounted at /home/node/.n8n, so it survives container restarts

On a Linux server without a desktop, install Docker Engine and Docker Compose as separate packages. Use Docker Compose for production, so n8n, PostgreSQL and a reverse proxy are defined together.

Installing n8n with npm is still possible, but n8n's documentation says the npm method is deprecated from n8n 3.0. For a new setup, use Docker.

When you need queue mode

A single n8n instance handles a lot. Queue mode is for when it doesn't: many executions at once, long-running jobs blocking others, or a need for more reliability.

Queue mode splits n8n into a main instance and one or more workers. It requires:

  • Redis, as the message broker
  • PostgreSQL, because n8n doesn't support running queue mode on SQLite
  • The same encryption key on the main instance and every worker

Most small businesses and agencies don't need queue mode at first. Start with one instance on PostgreSQL so moving later is simple.

These are starting points, not guarantees. Watch memory for the first few weeks and adjust.

Want a quick answer for your setup? The free n8n server size calculator asks five questions and suggests a starting size.

n8n system requirements by workload: recommended vCPU, RAM and database for each server size
Recommended n8n server sizes by workload.
Workload Example Server to start with
Testing and learning A few manual workflows 1 vCPU, 1 to 2 GB RAM, SQLite
Light production Lead alerts, form to CRM, daily reports 1 to 2 vCPU, 2 GB RAM, PostgreSQL
Agency or multi-client Many scheduled and webhook workflows 2 vCPU, 4 GB RAM, PostgreSQL
Files, AI and large data PDFs, images, AI agents, big syncs 2 to 4 vCPU, 4 to 8 GB RAM, PostgreSQL
High volume Thousands of executions an hour Queue mode with Redis, PostgreSQL and workers

AI agent workflows mostly call external models, so the model doesn't run on your server. Their prompts and responses still add to the data each execution holds in memory.

Self-hosted n8n or n8n Cloud?

Self-hosting gives you control and a lower cost at volume, but you're responsible for updates, security, backups and monitoring. n8n Cloud handles the server for you. To compare the Cloud plans with the free self-hosted edition, see our n8n pricing 2026 breakdown.

If a workflow touches client leads or patient enquiries, someone has to notice when it fails. If you self-host, set up these from day one:

  • Daily database backups, plus a restore you have actually tested
  • HTTPS for the editor and webhooks
  • An error workflow that alerts you when something fails
  • An uptime check on the instance

Many agencies run n8n next to an all-in-one CRM. If you are weighing one, our GoHighLevel pros and cons review explains where each tool fits.

The short answer

Start with 1 to 2 vCPUs, 2 GB of RAM, an SSD and PostgreSQL, running in Docker behind HTTPS. Move to 4 GB or more if you process files or large datasets. Add queue mode with Redis only when one instance can no longer keep up.

Sources: n8n Docker installation, n8n memory issues, n8n database options, n8n queue mode, n8n sizing prerequisites.

Frequently asked questions

What are the minimum system requirements for n8n?

n8n's own documentation gives illustrative figures of 320 MB to 2 GB of memory and 512 MB to 4 GB of SSD storage, and notes that an idle n8n Cloud instance uses about 100 MB of memory. For a real self-hosted server, 1 vCPU and 2 GB of RAM is a sensible starting point for light workflows.

How much RAM does n8n need?

It depends on your workflows more than on n8n itself. Large JSON payloads, binary files, the Code node, manual test runs and many workflows running at once all raise memory use. Plan for 2 GB for light use and 4 GB or more if you process files or large datasets.

Does n8n need PostgreSQL?

No. n8n uses SQLite by default. PostgreSQL is optional for a single instance, but queue mode does not support SQLite, so you need PostgreSQL if you plan to scale with workers.

Can I run n8n on a Raspberry Pi or an old PC?

For personal or test workflows, yes, as long as it can run Docker and has enough memory for your workflows. For client work that must run reliably, use a server with backups and monitoring.

What port does n8n use?

The official Docker setup maps n8n to port 5678. In production, put a reverse proxy in front of it so the editor and webhooks are served over HTTPS.

Get in touch

Have a question about my work or something I wrote? Send me an email.