Skip to content
Servers

How Much VPS Do You Need for n8n? A Practical Hosting Guide

In this article8 sections

A small n8n instance does not need a large server, but the right VPS depends on what your workflows do. Light automations that mostly call APIs can start on a modest plan; workflows that process large files, run several jobs at once, or keep a long execution history need more memory and storage. Think about the workload first, then choose a server with room to grow.

This guide helps you estimate what to run, explains the parts of an n8n deployment that affect performance, and gives you a checklist for keeping the instance reachable and backed up. If you want a server with n8n, PostgreSQL, Nginx and SSL installed at handover, see Siteharbour’s n8n VPS hosting. For an existing setup where webhooks fail behind Nginx, use our n8n webhook troubleshooting guide.

Start with the work your workflows will do

There is no single RAM number that suits every n8n instance. The editor and basic workflows are only part of the load. Each active execution can hold data in memory while nodes process it, and one workflow that downloads a large file can use more resources than many small API calls.

Use these as planning bands, not guaranteed minimums:

Workload What to start with What to watch
Learning or a few personal workflows Small VPS with 2 vCPU and around 4 GB RAM Memory during executions; disk space for history
Small production instance with several active workflows At least 4 vCPU and 8 GB RAM gives more room for concurrent work Peak RAM, database size and workflow duration
Large files, frequent jobs or several users Size from measured execution load; consider more memory and faster storage Concurrent executions, binary data and database I/O
Local AI models on the same server Size for the model separately; CPU inference can be slow and memory hungry Model memory, context size and simultaneous workflows

n8n does not reserve a fixed block of RAM per workflow. Measure an actual run that resembles your busiest work before choosing a production size. Keep spare memory for the operating system, database, reverse proxy and update jobs.

Estimate CPU, memory and storage separately

CPU matters when workflows transform data, run code, parse documents or execute many jobs at once. If most nodes wait on remote APIs, adding cores may not solve slow responses. Look at the time spent in each node before scaling up.

Memory is often the first limit to hit. File downloads, large JSON responses, image processing and parallel executions can all raise the peak. If the host starts swapping, workflows may become slow even if the average CPU graph looks quiet.

Storage holds the operating system, n8n data, database, logs, workflow history and binary files. Set execution retention deliberately. Keeping every successful execution forever can make a small disk fill up even when the workflows themselves are stable.

For a small production setup, use PostgreSQL rather than treating a database choice as an afterthought. Keep the data directory and database on persistent storage, and include both in a backup plan.

Cloud n8n or n8n on your own VPS?

Hosted n8n is a good fit when you want someone else to maintain the server and do not need control over the underlying machine. A VPS makes sense when you need to choose the server location, control the network and data, connect to private services, or run an instance with your own infrastructure practices.

Self-hosting also means owning updates, backups, uptime monitoring and access control. The software licence and paid feature availability can change, so check the current n8n documentation before building a plan around a specific feature. Do not compare cloud and self-hosted options only by monthly server price; include the time it takes to keep the instance healthy.

What a reliable n8n VPS setup includes

  1. A supported Linux release and a maintained n8n version. Pin and update the deployment deliberately, rather than letting an unattended image change the application without a rollback plan.
  2. A public domain with HTTPS. Webhook callers need a stable public URL. Put a reverse proxy such as Nginx in front of n8n and configure the public webhook URL so generated links use the correct HTTPS address.
  3. Persistent application data. Mount the n8n data directory outside an ephemeral container layer. Otherwise, replacing the container can remove the configuration and encryption key.
  4. A stable encryption key. Back it up separately. Credentials are encrypted with it; losing it can make saved credentials unusable after a restore.
  5. Authentication and a narrow firewall. Expose only the ports you need, keep the editor behind login, and do not publish the internal n8n port directly to the internet when a reverse proxy is used.
  6. Database and file backups. Schedule backups outside the VPS and check that they can be restored. A snapshot alone may not be a consistent database backup.
  7. Execution pruning and monitoring. Retain the execution history you need for support, then remove old data. Watch disk, memory, failed jobs and webhook response times.

How to tell when the VPS needs more resources

Use the graphs from your provider and n8n execution history together. A recurring memory peak during a file workflow points to a different fix than a CPU spike when many jobs run at once. Before upgrading, check whether the workflow can avoid loading a large file into memory, reduce unnecessary parallel runs, or prune old execution data.

  • Memory repeatedly approaches the plan limit: reduce the size or concurrency of jobs, or add RAM.
  • Disk fills steadily: inspect execution retention, database growth, logs and uploaded files.
  • CPU stays high during specific workflows: profile those nodes and check whether they process more items than expected.
  • Webhooks are late but the server is quiet: check DNS, reverse proxy timeouts and the remote service that triggers them.
  • Backups or upgrades fail: check free disk and memory during those operations, not only during ordinary traffic.

Keep a short note of the normal peak after launch. That gives you a baseline for future growth and helps distinguish a real capacity issue from one slow third-party API.

Before you put business workflows on it

Make a test workflow that receives a webhook, stores a small record, and sends a response. Confirm that the production URL is HTTPS and works from outside your network. Then test a backup and restore on a staging copy before relying on the server for important customer or internal processes.

Keep credentials out of screenshots and support requests. Limit who can log in, review community nodes before installing them, and remove access that a contractor no longer needs. An automation server often has credentials for many other services, so its access deserves the same care as an application server.

Choose the next step

If you already administer Linux servers, compare a self-managed plan with your estimated peak load and use the setup checklist above. If you want the webhooks, database and HTTPS configuration handled as part of the initial setup, compare the n8n VPS service and ask for a plan sized to your workflow volume. For a general look at server sizing and access, see Siteharbour VPS plans.

Frequently asked questions

Is 4 GB RAM enough for n8n?

It can be enough for a light single-user instance with small API-based workflows. Large files, parallel runs and local AI models can need substantially more. Check actual peak memory before using a small plan for production.

Does n8n need a VPS?

No. n8n offers a hosted option and can also run on other infrastructure. A VPS is useful when you want control over the server, networking, storage location or access to private systems.

Should I use Docker for n8n?

Docker is a common way to run n8n because it keeps the application environment predictable. Use persistent volumes, a stable encryption key, a database backup and a planned update process; a container by itself does not provide those.

Why do n8n webhooks show localhost?

The instance may not know its public HTTPS address. Set the webhook URL and proxy settings for the domain, then check the reverse proxy configuration. The Nginx webhook guide covers the common settings.

Share this guide