What this guide adds
This guide puts the measuring tools and the price data in one sequence of work: size the stack, turn it into a server, add traffic and region, then read the plans of the VPS Price Index next to the result. The /cloud/vps page lists what to check; this one shows how to measure each check.
Before you compare
Most VPS choices go wrong in the same place: a plan is picked first and the apps are made to fit it. This guide goes the other way. You measure what the apps need, add headroom and traffic, pick a region from your own latency, and only then compare plans.
Every figure on this page comes from the site's data, with its source and the day it was read. The text itself states no prices, and no provider is ranked.
What to compare
- Memory first
- Memory is what self-hosted stacks run out of first. A plan with more vCPUs and too little memory swaps or kills processes; size memory from the apps, then check the CPU.
- Shared or dedicated CPU
- Shared vCPUs suit light load that comes in bursts. Steady work, such as builds, transcoding or a busy database, is where dedicated cores make a difference. Read what the plan says about the CPU it gives you.
- Disk and how it grows
- Data, logs and container images grow. Check the disk size, whether block storage can be attached later, and whether snapshots are billed separately.
- Transfer and egress
- Some plans include a transfer allowance and bill past it; others bill all egress. Measure your traffic before you read the allowance, and count what leaving would cost.
- Region
- Choose the region nearest the people who use the apps, measured from where they are, and check which law the data falls under there.
- A way out
- Check that disks and data can be exported, and plan for a new IP address when you move.
Measure with the tools
In this sequence. Each step is answered by a free tool on this site that works in your browser.
- Size the stack
Pick the apps you will run and how many people use them. The Stack Sizer gives memory, CPU and disk from each app's official requirements, with the source of every figure.
Tool: Stack Sizer
- Turn it into a server size
For a workload outside the catalog, or to add peak load and data growth, the server size calculator gives a vCPU, memory and disk range with a breakdown you can change.
Tool: Server size calculator
- Measure the traffic
Enter page views, downloads and API calls. You get the monthly transfer and the peak bit rate, to compare with the plan's transfer allowance and port speed.
- Price the egress and the exit
Enter what you send to the internet and what you would move out if you left. You get both costs at each provider's dated list prices, so a plan that looks cheaper is checked against its traffic.
Tool: Cloud egress calculator
- Pick the region from your own latency
Run the latency test from where your users are. Your browser times requests to each provider's published test endpoint; nothing goes through this site.
Tool: Latency test
Dated prices from the Price Index
Plans from the VPS Price Index sample, as listed in the dataset, with the vCPUs and memory of each. Read a row only after the steps above have told you how much memory you need; the full index has every sampled size, normalized per vCPU and per GiB.
| Provider | Plan | Region | vCPU | RAM (GiB) | Listed price per month |
|---|---|---|---|---|---|
| DigitalOcean | Basic Droplet (Regular) 4 GiB | One price everywhere | 2 | 4 | $24 |
| DigitalOcean | Basic Droplet (Regular) 8 GiB | One price everywhere | 4 | 8 | $48 |
| Akamai Cloud | Linode 4 GB (Shared CPU) | North America | 2 | 4 | $24 |
| Akamai Cloud | Linode 8 GB (Shared CPU) | North America | 4 | 8 | $48 |
| AWS | Lightsail 4 GB (Linux/Unix, IPv4) | One price everywhere | 2 | 4 | $24 |
| AWS | Lightsail 8 GB (Linux/Unix, IPv4) | One price everywhere | 2 | 8 | $44 |
| UpCloud | Starter 4 GB, 2 cores | One price everywhere | 2 | 4 | €12 |
| UpCloud | Starter 8 GB, 4 cores | One price everywhere | 4 | 8 | €20 |
Prices as each provider lists them, in its currency, excluding taxes unless the row's source says otherwise. The full index converts them with dated rates.
The full index: VPS price index
Common mistakes
- Choosing by vCPU count
- Two plans with the same vCPU count can give very different CPU time: one shared, one dedicated. Read what the plan says about the CPU, not only the count.
- Forgetting the traffic bill
- A low plan price with egress billed past a small allowance can end up costing more than a larger plan with transfer included. Price the traffic with the egress calculator.
- No room to grow
- A server sized exactly for today runs out on the first busy day. Keep headroom for peaks and for data, as the sizing tools do.
- Backups next to the server
- A snapshot or backup kept with the same provider, in the same account, is not a way out. Keep one copy elsewhere and test a restore.
- Converted prices taken as exact
- Plans are listed in different currencies. The index converts them with dated reference rates; your card or invoice uses its own rate and taxes.
Limits
- The sample covers a few providers and plan sizes, not the market. A provider missing here is not a judgment about it.
- Prices change: every row shows the day it was read and is marked stale after the refresh window. The provider's page is authoritative.
- Performance was not measured: plans with the same figures can behave differently under load.
On the /cloud page
The /cloud page for this kind of compute lists what to check before you commit, and describes the principles ANAROQ is building to, with its In development status.