Skip to content
ENRO
Enter the Universe
Free tool · VPS

VPS size calculator

Tick what you want to run and how much traffic you expect. The calculator adds up typical needs, keeps room for the operating system and picks the smallest VPS that fits.

What will run on the VPS?
Traffic
You need about0.8 vCPU · 1.9 GB RAM · 13 GBRecommended plan

Developer Compute · 1 vCPU · 2 GB RAM · 30 GB NVMe · €4.99/mo

Choose this VPS

Includes the operating system and 25 % RAM headroom. An estimate — real use depends on your code and visitors.

How it estimates

Every workload has a typical footprint: a small Discord bot needs a few hundred MB of RAM, a WordPress site with its database around 1 GB, a database server or a Docker stack 2 GB or more. The calculator adds them, adds the operating system, keeps 25 % RAM headroom and scales CPU with traffic. Disk is the sum of what the workloads typically use.

Signs you picked too small

  • Memory is always above 90 % and the system starts swapping.
  • Load average stays above the number of vCPUs.
  • Builds or database queries that were fast become slow at busy hours.

Start small, measure, then grow

Most projects start on less than people think. Run htop or your monitoring for a week and look at peak use, not averages. Because billing is monthly, you can move to a bigger plan when the numbers say so instead of paying for capacity you do not use.

Frequently asked questions

How much RAM does a VPS need for a website?
A small static or PHP site runs in 512 MB–1 GB. WordPress with MySQL is comfortable from 2 GB. Several sites, or sites with heavy plugins, need 4 GB.
Is 1 vCPU enough?
For a bot, a small site or a VPN, yes. Databases, builds, game servers and busy apps benefit from 2–4 vCPU.
How much disk space do I need?
The OS takes about 8 GB. Add your apps, databases, Docker images and logs. Backups should live somewhere else, not only on the same disk.
Can I run several things on one VPS?
Yes — that is what the calculator adds up. Docker makes it easier to keep them separate. Split them onto two servers when one workload's spikes start slowing the others.