Performance guide · Edinburgh
Edinburgh: the August load test you did not schedule
Once a year the city doubles and every venue, restaurant and letting site finds out what its hosting is made of. The instrumentation has to be in place before that, not during it.
The short answer
Plan for the spike by widening cache coverage first: a cached page consumes almost no CPU, so the number of visitors a plan can absorb depends far more on your hit ratio than on the tier you bought. Then instrument, so that when August arrives you are reading numbers instead of guessing.
HostingFast runs NVMe storage behind LiteSpeed caching from London with free migration and free SSL, and every account states its CPU and memory allocation — so capacity planning is arithmetic rather than optimism.
By the HostingFast team · Reviewed 15 August 2026
Free
Migration, engineers do it
24/7
Cover through the peak
Daily
Backup cadence
$0
Setup charge
The traffic curve here is flat and professional for most of the year. Then the festival month arrives, a site that was comfortable at forty concurrent visitors meets several hundred, and nothing in the code has changed to explain why it fell over.
That gives the city an unusual hosting profile: eleven months of steady professional traffic and one month of something closer to a launch. Everything below is ordered around surviving the second without over-buying for the first.
What actually runs out when the traffic arrives
Not disk and rarely bandwidth. The ceiling is concurrent PHP execution against the CPU and memory the plan allocates — one core and 2 GB on the entry and mid shared tiers, two cores and 4 GB at the top. Each uncached request holds a worker for as long as it takes to build the page.
Cached requests barely register against that budget. Which is why a site with broad cache coverage absorbs a spike on the same plan that a poorly cached site falls over on, and why the first capacity decision is always a caching decision.
Instrument in June, not on the first Friday of August
Put an external monitor on an uncacheable URL so an outage cannot hide behind a warm cache. Record a baseline TTFB and LCP for the pages that will carry the traffic — programme, listings, booking — and keep the numbers where you can compare against them under load.
Then rehearse the failure. Generate load against a staging copy with a simple tool such as k6 or ab, watch which page degrades first, and fix that one. Discovering it in front of a queue of ticket buyers is considerably more expensive.
The pages a spike cannot cache
Booking flows, baskets, account areas and anything behind a login are uncacheable by design, and they are exactly the pages a festival audience hammers. Cut their query count, add the indexes the reports need, and make sure a third-party availability lookup cannot block the render.
If one of those pages is genuinely the bottleneck after the query work is done, that is the moment to move up a tier or onto a VPS. Plan changes happen in place from the client area with no migration and no downtime, so the decision can wait until the evidence is in.
Migrating without a measurable regression
Never migrate in your busy month, and never migrate without a control. Our engineers copy the site across free while the old host keeps serving, which leaves two live copies you can measure with the same tool from the same connection.
Compare the pair, sign it off, then move delegation. Afterwards, re-run the identical measurements — a migration that improved nothing is a fact worth knowing, and one that quietly regressed a template is worth catching in the week rather than in August.
Capacity you can reason about
Every tier states its CPU and RAM allocation, storage is NVMe throughout, and LiteSpeed caching runs in the web server rather than as a plugin. That means the capacity conversation is about numbers you can check rather than adjectives on a comparison page.
Setup costs nothing, migration costs nothing, and the rate you sign up at is the rate you renew at — so scaling up in August and back down afterwards is a plan change, not a negotiation.
- .co.uk or .scot registered and renewed with the plan
- Stated CPU and RAM on every tier
- In-place upgrades, no migration when you change plan
- Free migration with the old site left serving
Why HostingFast
Standard on every plan
Cache-first capacity
A LiteSpeed hit is answered before PHP loads, so widening coverage raises the number of visitors a plan absorbs without changing the plan.
Allocations you can plan against
One core and 2 GB on the entry and mid shared tiers, two cores and 4 GB at the top. Stated, not implied.
Upgrade in place
Plan changes run from the client area with no migration and no downtime, so a decision made on Thursday is live on Thursday.
A control for every migration
The old site keeps serving while the copy runs here, so a regression is caught by comparison rather than by a customer.
Backups on a daily cadence
Every plan backs up daily and restores from the panel in minutes — files, databases or both, without a ticket in the way.
Support through the night
First reply inside two hours, and inside one when a site is down, at whatever hour it happens. In a month where the traffic peak lands at 22:40 rather than 10:00, that is the difference that matters.
Quick Start
From order to online
- 1
Widen the cache
Find the templates that are missing and fix the exclusions, cookies and query strings causing it. This is the cheapest capacity you will ever buy.
- 2
Instrument the uncacheable pages
Baseline TTFB on booking, basket and account screens, and put an external monitor on one of them.
- 3
Rehearse on staging
Push load at a clone, note which page degrades first, fix it, repeat. Better a bad afternoon in June than a bad night in August.
- 4
Decide the tier on evidence
If the numbers still say you are short after the query work, upgrade in place. If they do not, keep the money.
Built In
Loaded onto every plan
- Cache coverage checked template by template, not site-wide
- External monitoring pointed at a page that cannot be cached
- A rehearsed load test against a staging clone before the busy month
- Stated CPU and RAM allocation on the plan you are buying
- In-place account upgrades — no migration when you change plan
- Daily backups, with self-service restores you run from the panel
- Free migration run by our engineers, with the old site left serving
- Human support on duty every hour of every day
- No setup charge, and a renewal that matches the first term
- Money-back cover: 30 days on hosting plans, 7 on reseller
Frequently Asked
What people ask us most often
How do I know whether my plan will survive August?
Test it rather than estimate it. Clone to staging, push load with k6 or ab at roughly the concurrency you expect, and watch which page degrades first. If the failure is on cached pages you have a caching problem; if it is on booking and basket pages you have a query problem or a genuine capacity problem, and only then is a bigger tier the answer.
Can I move up for the busy month and back down afterwards?
Plan changes happen in place from the client area with no migration and no downtime, so moving up before the season is straightforward. Do the arithmetic on the annual terms before you decide, and remember that cache coverage is usually cheaper than capacity.
Should we migrate hosts before or after the festival?
After, and never during. Migrations here are free and run with your old site still serving, which means you can prepare and compare at any time — but the delegation change is the one irreversible-feeling moment in the process, and it should not share a week with your busiest trading of the year.
Why does our booking page slow down when nothing else does?
Because it cannot be cached and everything else can. Each booking request runs the full PHP and database path, often with a third-party availability call inside it. Count the queries, add the missing indexes, and make sure the external lookup has a timeout that cannot hold the render hostage.
Keep reading
Hosting in the United Kingdom
Rates in GBP, which domain endings we can register there, and how the tax works out.
Web Hosting in Cork
Irish pharma and food sites, and the round trip from Cork to a London origin.
Web Hosting in Copenhagen
Danish audiences, .dk, and a market with very little patience for a slow page.
CDN (Edge Delivery)
What an edge network caches for you, and the requests it will never touch.
VPS Hosting
KVM virtual servers with root access and CPU you are not sharing when August lands.
Website Builder
A hosted builder for the sites that do not need a codebase behind them.
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.
Be ready before the traffic is.
Stated allocations, in-place upgrades and free migration with a control copy — so the busy month is arithmetic rather than nerves.
View VPS Hosting plans