How much RAM does my VPS need?

There is no single figure that fits every VPS, because it depends entirely on what the server is running — the operating system itself is a small, fairly fixed cost, and everything on top of it is what actually varies. If you already have a server running something similar to what you plan to deploy, free -h will give you a far more accurate answer than any general rule.

What actually consumes memory

• The operating system and its base services — typically a few hundred MB, fairly consistent regardless of what else you run.
• A database server — MariaDB or MySQL can be tuned to use very little or a great deal, depending on its buffer settings and how much data it is working with.
• A web server and its worker processes — Nginx is generally lighter than Apache under equivalent load, but the bigger factor is usually what is generating the dynamic content behind either one.
• PHP-FPM or an application runtime — each worker process holds its own memory, so the number of concurrent workers matters more than any single request's footprint.
• Anything else you install — a caching layer, a queue, a second application — each adds its own baseline on top of everything above.

Measure it on a running server

free -h

Read the available column rather than free — Linux uses spare memory for disk caching, so a low free figure with a healthy available figure is completely normal, not a sign of a problem. See seeing what is using your server resources for the fuller picture, including which specific process is consuming what.

Rough starting points

These are directional, not specifications — for the actual RAM tiers we offer and what each is generally suited to, see /vps-hosting.

Signs you have genuinely outgrown your current tier

• free -h consistently shows very little available memory, not just during occasional spikes.
• Swap usage is climbing steadily rather than sitting near zero — see adding swap space, which is a short-term cushion, not a fix for this.
• A service is being killed outright by the kernel's out-of-memory handler — check dmesg or the system log for Out of memory: Killed process.
• Performance degrades specifically under load rather than being uniformly slow, which points at resource exhaustion rather than a configuration problem.

If you are seeing several of these together on a server that has not changed recently, that is a genuine capacity signal rather than something to tune your way out of indefinitely. If your workload might be outgrowing a VPS altogether rather than just needing a bigger tier of one, VPS or dedicated server: where the line is covers that decision, and our team can help you size a move either way.