Layer 5 · self-hosting reality check
What it actually takes to self-host n8n
The docs say 512 MB. In practice you want 2 GB, and 4 GB once workflows run concurrently. Here is the honest version — real requirements, real monthly cost, what you will be maintaining, and the one thing that catches people out.
Usually reached from Zapier alternatives, where n8n is one of the picks.
| RAM — documented minimum | 512 MB |
|---|---|
| RAM — what it really needs | 2 GB, and 4 GB once workflows run concurrently |
| CPU | 2 vCPU |
| Disk | 20 GB, mostly execution history |
| Monthly cost | $12–25/mo on a small VPS |
| Setup time | 45 minutes |
| How you install it | docker compose with PostgreSQL — do not leave it on the default SQLite |
| Ongoing maintenance | Active. n8n ships fast and node changes occasionally break workflows. |
| Where it stops scaling | Thousands of executions a day on one node; queue mode with workers goes considerably further. |
The thing that catches people out
Execution history is stored in the database and n8n keeps it forever by default. A busy instance quietly grows to tens of gigabytes and then slows to a crawl, because every workflow list query walks that table. Set `EXECUTIONS_DATA_MAX_AGE` on day one; almost nobody does, and the symptom appears months later as 'n8n got slow'.
When not to self-host n8n
Your automations are business-critical and you have no on-call. A workflow engine that is down is an outage in every system it touches.
Every guide here carries this section. A site that only ever tells you to self-host is selling something — the useful answer is sometimes no.
Other Layer 5 self-hosting guides
- Self-hosting LibreOffice2 GB with a large spreadsheet open
- Self-hosting ONLYOFFICE6 GB for the Document Server with a handful of concurrent editors
- Self-hosting Collabora Online4 GB, and roughly 1 GB per 20 concurrent documents
- Self-hosting CryptPad2 GB for a small instance
- Self-hosting Mattermost4 GB for a team of 50 with PostgreSQL on the same box
- Self-hosting Rocket.Chat6 GB with MongoDB on the same machine
Common questions
- How much RAM does n8n actually need?
- 2 GB, and 4 GB once workflows run concurrently in practice. The documented minimum is 512 MB, which is the figure at which the process starts rather than the figure at which it works under real use. 2 vCPU alongside it.
- What does self-hosting n8n cost per month?
- $12–25/mo on a small VPS This is commodity VPS pricing and excludes your time, which is the larger cost for most people — budget for active. n8n ships fast and node changes occasionally break workflows.
- How long does it take to set up n8n?
- 45 minutes, via docker compose with PostgreSQL — do not leave it on the default SQLite.
- When should I NOT self-host n8n?
- Your automations are business-critical and you have no on-call. A workflow engine that is down is an outage in every system it touches.
- What is the most common mistake when self-hosting n8n?
- Execution history is stored in the database and n8n keeps it forever by default. A busy instance quietly grows to tens of gigabytes and then slows to a crawl, because every workflow list query walks that table. Set `EXECUTIONS_DATA_MAX_AGE` on day one; almost nobody does, and the symptom appears months later as 'n8n got slow'.