The two questions every VPS buyer asks are the same: how much RAM, and how much storage? Both are easy to get wrong, and the mistakes cost differently: oversize and you pay monthly for capacity you never touch; undersize and you discover it during a traffic spike, when memory pressure slows every request and the disk fills with logs.
Sizing depends on your workload, not on marketing tables — but with structure, you can estimate both numbers in ten minutes. This guide explains what actually consumes RAM and NVMe storage, then maps the results to real plans.
What Actually Consumes RAM
RAM is the resource your software notices first. When memory runs out, the operating system starts swapping to disk, and every operation slows dramatically. The main consumers on a typical VPS:
- The operating system and services. A modern Linux distribution, web server, database and monitoring agent together occupy roughly 300–600 MB before your application starts.
- Your runtime and application. PHP-FPM workers, Node.js processes or Java heaps each reserve memory per process. Ten PHP-FPM workers at 80–120 MB each is close to a gigabyte on its own.
- The database. MySQL or MariaDB caches data and indexes in memory by design — more RAM means fewer disk reads. This is often the difference between a snappy site and a sluggish one.
- Caches. Redis, Memcached, OPcache and page caches are memory-hungry by choice: that is their job. A Redis object cache is happy to use several hundred MB.
- Concurrency. More simultaneous visitors, more parallel cron jobs, and build processes all multiply the per-process footprint.
The practical consequence: a single small application fits in 1–2 GB, a WordPress site with caching and a database in 2–3 GB, and anything running several services, a bigger database or builds will pull 4 GB and up.
What Actually Consumes NVMe Storage
Storage is easier to estimate because it grows in predictable ways:
- The operating system and base software occupy several gigabytes before your data exists.
- Application code and dependencies — usually small, but container images and package caches add up quickly.
- Databases grow with usage; log and binary files inside the database directory count too.
- Uploads and media grow most on content-heavy sites.
- Logs and temporary files grow silently; a misconfigured log can fill a disk in days.
- Backup and update headroom — every update downloads packages, and restore operations need space to work in.
A working rule: plan for today's data, expected growth over 12–24 months, plus 20–30% of free headroom for logs and updates. Storage is billed monthly, so idle space is waste — and so is running out.
Sizing by Workload: The Open Vista Lineup
Open Vista, an IT consulting firm based in Rabat, Morocco, offers five NVMe VPS plans in US dollars per month, with no contract. Every plan includes a dedicated IPv4 address, a management panel, automatic backups, 24/7 support in English and French, and deployment in under 15 minutes.
| Plan | Price (USD/month) | vCPU | RAM | NVMe storage |
|---|---|---|---|---|
| VPS-1G | $9.99 | 1 | 1 GB | 30 GB |
| VPS-2G | $19.99 | 1 | 2 GB | 40 GB |
| VPS-3G | $34.99 | 2 | 3 GB | 50 GB |
| VPS-4G | $39.99 | 2 | 4 GB | 80 GB |
| VPS-5G | $49.99 | 3 | 5 GB | 100 GB |
How the plans map to workloads:
- VPS-1G (1 vCPU / 1 GB / 30 GB) — one small application, a lean site with caching, a light API, a staging box or a VPN endpoint — realistic only if you keep the database small.
- VPS-2G (1 vCPU / 2 GB / 40 GB) — the practical entry point for one production WordPress site or a small application plus its database.
- VPS-3G (2 vCPU / 3 GB / 50 GB) — the most popular plan: two vCPUs and 3 GB run a typical small production application, a WordPress store or a small team's development server.
- VPS-4G (2 vCPU / 4 GB / 80 GB) — several applications or a database with real cache needs, plus background jobs.
- VPS-5G (3 vCPU / 5 GB / 100 GB) — the largest standard plan: more concurrency, larger databases, media-heavy sites or CI runners.
When comparing cheap VPS offers, price is only one axis: check for a dedicated IPv4, real NVMe storage, automatic backups and reachable support — the VPS-3G's popularity reflects that balance, not just its price.
Start Right-Sized and Grow Later
Two habits keep sizing under control:
- Measure, then decide. Run the application on the smallest plan you believe fits, watch `free -h`, disk usage and response times for a week, and upgrade only when the data says so. Guessing upward wastes money; guessing downward and ignoring the signs wastes time.
- Respect your growth. Applications grow: databases, media and logs do not shrink on their own. A plan that fits today with 25% headroom is better positioned than one sized to the exact current usage.
Because Open Vista bills month to month with no contract, sizing is a reversible decision. You can start at VPS-2G, confirm the behaviour, and move to VPS-3G or VPS-4G when the metrics justify it — and automatic backups make each move safer. Geography is a separate axis: if your users are mainly in the USA or Europe, latency follows network path and caching more than RAM, so size by workload and treat delivery as a separate question.
Quick Sizing Tips
- Count your services: OS + database + runtime + cache is your baseline before users arrive.
- Add 20–30% free disk headroom for logs, updates and growth.
- For content sites, budget uploads and media explicitly — they keep growing.
- Upgrade when memory consistently sits above 85% or the disk above 75%, not when the site already crashed.
- With monthly billing, right-size down as well as up when a project ends.
FAQ
Is 1 GB of RAM enough for a VPS in 2026? For a single lean application with caching enabled, yes — the VPS-1G covers a light site, API or staging environment. For anything with a real database, several services or expected traffic, 2–3 GB removes most memory anxiety for a modest price difference.
What consumes NVMe storage the most? Media uploads, growing databases, container images and unchecked logs. Database and log growth is easy to underestimate because it is silent — which is why regular disk monitoring belongs in your maintenance routine.
I need a VPS whose users are mainly in the USA or Europe. Does that change the specs? Latency for users in the USA or Europe is driven by network path, routing and content delivery, not by the server's RAM. Size the plan by workload and treat delivery speed as a separate question — caching and a CDN usually do more for distant visitors than extra memory.
Which plan do most customers choose, and why? The VPS-3G at $34.99 per month — two vCPUs, 3 GB of RAM and 50 GB of NVMe storage — is the most popular plan because it covers a typical small production application or a WordPress site with room to grow, at a price that stays predictable month to month.
Right-sizing starts with real numbers, not speculation. Compare the five Open Vista NVMe VPS plans — no contract, dedicated IPv4 on every plan, automatic backups and 24/7 support in English and French — at https://openvista.ma/en/vps, and order the one that fits your workload at my.openvista.ma.