Capacity Notes
Shared vs VPS hosting — Nobody should buy a VPS without a graph to point at
Something on your site is slow, a forum told you to move to a VPS, and nothing you have looked at so far distinguishes a genuine resource ceiling from an unconfigured cache.
The short answer
Look at the numbers before you look at the plans: if your cPanel resource-usage graphs are not repeatedly pinning entry processes, CPU or I/O over a full week, the thing slowing your site is a cache miss or a slow query and a VPS will not touch it.
A VPS earns its place when you need something shared hosting genuinely cannot run — root, a custom daemon, a non-standard stack — or when a real ceiling shows up in a week of data. Below, how to read that data, what to fix first, and what a KVM box actually hands you along with the root password.
By the HostingFast team · Reviewed 24 August 2026
Free
Migration, engineer-run
24/7
Answers at any hour
Daily
Restore point cadence
$0
Joining fee
The shared-versus-VPS argument is normally conducted entirely without evidence. One side says shared hosting is throttled, the other says a VPS is overkill, and neither has opened the resource graphs on the account being discussed.
There is a measurement that settles it, and it takes about ten minutes. Read the account's own usage history, compare a cached response with an uncached one, and check whether the database is the thing keeping PHP waiting. Whatever that produces, you will know which of the two you are buying and why.
Read the graphs the account already keeps
cPanel records resource usage per account and keeps a history. Open it and look for three specific things: entry processes hitting the limit, CPU pinned at the ceiling for sustained periods, and I/O throttling. A brief daily spike is normal. A flat line along the top for hours is a ceiling, and that is a different conversation.
If the graphs are quiet and the site still feels slow, the bottleneck is inside your application. That is good news, because application bottlenecks are cheap to fix and hardware ones are not.
Do this over a week, not an afternoon. Traffic patterns are weekly, backup jobs and cron runs land at fixed hours, and the load profile of a Tuesday morning tells you very little about the Sunday evening peak that was actually annoying your visitors.
Separate the cache from the hardware
Two curl runs settle most of this argument. Request a page as an anonymous visitor and read the total time; request it again with a cache-busting query string and read it again. The first number is what your visitors get from the LiteSpeed cache, the second is what PHP and MySQL cost when nothing is cached.
If the cached number is quick and the uncached number is slow, you have a PHP or database problem, not a plan problem. Turn on the slow query log, look at what is autoloading out of the options table, and check whether an object cache would keep the same queries from running twice.
If both numbers are slow, look at the response headers first. A page setting no-cache, or a plugin that inserts a session cookie on every request, will defeat a perfectly good server cache and produce exactly the symptoms people blame on shared hosting.
What a KVM VPS hands you, and what it takes back
Our VPS range is KVM, not a container, so the kernel is yours and root really is root. You choose from 24 Linux and BSD images, pick London or Dallas as the region, and sit behind more than 1 Tbps of DDoS filtering on hypervisors with a 20 Gbps uplink. Snapshots are on demand and cPanel can be bolted on if you want it.
What comes back the other way is a job list. Kernel and package patching, firewall and access rules, monitoring that actually reaches you, and periodic proof that your backups restore rather than merely exist. None of it is exotic, but all of it is now yours.
Worth knowing before you order: VPS and dedicated servers are built to order the moment payment lands, so they sit outside the refund window that covers shared, business, WordPress and WooCommerce plans.
The shared tier most people are actually looking for
Turbo is where the majority of this traffic lands: 20 GB of NVMe, 25 sites, a terabyte of bandwidth and 50 MySQL databases, with LiteSpeed in front and per-site PHP version switching in the panel. For a site whose graphs are not pinned, that is more headroom than the workload is using.
Upgrading is an account change rather than a move. Sprint, Turbo and Nitro sit on the same platform, and stepping up does not involve a migration, a new IP or a DNS change.
If the week of data does argue for a server, the same supplier still has it. That matters more than it sounds: the alternative is discovering the ceiling and then having to switch host and stack at the same time.
We would rather sell you the right tier
A customer who buys a VPS to fix a caching problem gets a slower site and a second job. That ends in a cancellation, so the incentive here points the same way the advice does.
The rate you order at is the rate you renew at, which also means an upgrade decision is never distorted by an introductory price expiring underneath you.
- Read a week of graphs, not an afternoon
- Cached and uncached, measured separately
- KVM, root, and the job list that follows
- Upgrades in place, no migration
Why HostingFast
Standard on every plan
LiteSpeed in front of every plan
Server-level caching rather than a plugin, so the cached path never wakes PHP in the first place.
Usage history you can actually read
cPanel keeps per-account resource graphs, which is the evidence this decision should be made on.
KVM virtualisation, not containers
Your own kernel, full root, 24 Linux and BSD images and a choice of London or Dallas.
DDoS filtering at the front
More than 1 Tbps of filtering ahead of the hypervisors, on a 20 Gbps uplink.
One platform, three tiers
Sprint, Turbo and Nitro upgrade in place from the client area — no migration, no new IP.
Flat renewals, every year
What you order at is what you renew at, so a capacity decision is never forced by a price rise.
Quick Start
From order to online
- 1
Collect a week of resource data
Open the resource usage history in cPanel and watch entry processes, CPU and I/O for seven days. Note when the peaks land and what was running at the time — a cron job and a traffic spike look identical on the graph until you check.
- 2
Measure cached against uncached
Curl the same URL twice, once plain and once with a cache-busting parameter, and compare the totals. A fast cached response and a slow uncached one is a database problem wearing a hardware costume.
- 3
Fix the cheapest cause first
Slow queries, autoloaded options and a cache defeated by a stray cookie cost nothing to fix. Only when those are clean and the graphs are still pinned does a server become the correct purchase.
Built In
Loaded onto every plan
- LiteSpeed caching built into the server, not patched in by a plugin
- NVMe SSD storage on every tier, including the entry plan
- Per-account resource graphs with history, so a ceiling is visible
- Per-site PHP version switching from the control panel
- Staging environments for testing a change before it ships
- Daily backups, with self-service restores you run yourself
- In-place upgrades between Sprint, Turbo and Nitro — no migration
- KVM VPS with full root, snapshots and an optional cPanel licence
- 1 Tbps+ DDoS filtering ahead of a 20 Gbps hypervisor uplink
- A renewal price identical to the price you ordered at
Frequently Asked
What people ask us most often
How do I tell a CPU limit from a slow query?
Look at where the time goes. A CPU or entry-process limit shows as throttling in the account's resource graphs and affects every request, cached or not. A slow query shows as a fast cached response and a slow uncached one, and it will be sitting in the slow query log with a duration next to it.
What Time to First Byte should a cached page return here?
Once LiteSpeed is serving the page, the origin work is close to nothing and the number is dominated by distance. From the UK and Ireland the indicative round trip to our London origin is about 5 to 15 milliseconds; from Western and Northern Europe, 10 to 35. If your cached TTFB is far above that band, the page is not being cached.
Does more RAM on a VPS make WordPress faster?
Only if something was swapping or a worker pool was starved. RAM that sits unused does nothing for response time. Check free memory under load first; if there is spare, extra RAM buys you nothing and the money would do more work as an object cache or a query index.
Is there a refund window if the VPS turns out to be wrong?
Not on VPS or dedicated servers — they are built to order the moment payment clears, so they fall outside the guarantee. Shared, business, WordPress and WooCommerce plans carry 30 days and reseller carries 7, which is another reason to prove the ceiling before you move off shared hosting.
Keep reading
NVMe Hosting Plans
Why the storage layer decides how a site feels between one request and the next.
Landing Page Hosting
Campaign pages that have to load instantly and cost almost nothing to run.
Shared Hosting (Glossary)
What shared hosting actually is at the kernel level, and which limits come with it.
VPS Hosting
KVM virtual servers with full root, DDoS filtering and one flat monthly rate.
Domain Names
Register or transfer a name, with the first year free alongside annual hosting.
Changing hosts? Run through our checklist first.
A straightforward sequence for a switch your visitors never feel: which files move first, how to shift email across without losing a single message, the right moment to repoint DNS, and the two mistakes behind almost all the downtime we get asked to rescue.
Buy the tier the data points at.
LiteSpeed on NVMe, resource graphs you can read, and in-place upgrades if the week of numbers says you need one.
View VPS Hosting plans