43 Minutes a Month: Why UK Firms Can’t Rely on Website Uptime SLAs
If they miss that target, you’re entitled to a service credit, but that credit won’t recover lost sales or a dented reputation. The percentage is a useful benchmark, not a guarantee your site will actually stay up.
TL;DR:
- A 99.9% uptime guarantee allows approximately 43 minutes of monthly downtime, but this measurement period significantly affects the perceived reliability.
- Uptime SLAs typically exclude scheduled maintenance, force majeure events, and issues caused by your own code or third-party integrations.
- Server-level uptime does not necessarily reflect end-to-end website availability, especially if external factors like DNS failures or API issues impact user experience.
- Independent monitoring of your site’s critical user journeys provides more accurate evidence of actual availability than relying solely on your provider’s reporting.
- A thorough SLA negotiation should specify measurement methods, response times, claim procedures, and include audit rights to verify uptime claims independently.
Table of Contents
What ‘uptime’ means and how to read percentage guarantees
Uptime is simply the proportion of time your website is available and working as expected. Downtime is everything else: outages, timeouts, and the moments your site fails to load for visitors. The tricky part is that a small difference in percentage hides a big difference in real time.
Uptime SLAs typically guarantee 99.5%, 99.9% or 99.99% availability, and providers listed on the Digital Marketplace commonly set their minimum around 99.5%. Here’s what each target means in practice over a month:
- 99.5% allows roughly 3 hours and 39 minutes of downtime a month
- 99.9% allows roughly 43 minutes of downtime a month
- 99.99% allows roughly 4 minutes of downtime a month
Whether that downtime is measured monthly or annually changes things too. A provider measuring annually can have a terrible week and still land within target once the rest of the year averages out, so always check the measurement period before comparing two SLAs side by side.
What a typical uptime SLA contains
An SLA is only as useful as the detail behind it. Most contracts define three things: how uptime is measured, what’s excluded from the calculation, and what you get when the provider falls short.
Monitoring usually relies on automated probes checking your site at set intervals from one or more locations. According to Gov, the real operational value of an SLA comes from clearly defined response and resolution times and escalation paths, not the headline percentage alone.
Common exclusions include:
- Scheduled maintenance windows, agreed in advance
- Force majeure events outside the provider’s control
- Incidents caused by the customer’s own code, configuration or third-party integrations
When a breach does occur, the claims process generally works like this:
- You notice or are notified of the outage
- You submit a claim within a set window, often just a few days
- The provider verifies the outage against their own monitoring
- A service credit is applied, usually capped as a percentage of your monthly or annual fee
Public-sector hosting agreements commonly define uptime availability at 99.9% while explicitly excluding scheduled maintenance from that figure, so read the exclusions before you trust the headline number.
Why an uptime percentage alone can be misleading
Here’s the gap that catches business owners out: your hosting provider’s SLA covers the infrastructure they control, not the whole journey a visitor takes through your site. Your server can be up while your website is still broken.
The NCSC’s guidance on Web Check makes the point plainly: server-level uptime is not the same as end-to-end availability. A visitor doesn’t care whether the server responded if the checkout page won’t load because of a failed payment gateway.
Outages that sit entirely outside a hosting SLA include:
- DNS failures that stop visitors reaching your domain at all
- CMS or plugin errors that crash specific pages
- Database faults that break dynamic content or forms
- Third-party API failures, from payment processors to booking widgets
Pro Tip: Map every external service your website depends on, then test the journeys that matter most, like checkout or contact forms, from a visitor’s point of view rather than just pinging the homepage.
How to measure and monitor your website’s real availability
You don’t need to take your provider’s word for it. Independent monitoring gives you your own evidence, which matters both for peace of mind and for backing up any SLA claim.
- Set up external monitoring from multiple probe locations, not just one, so a regional network issue doesn’t get missed
- Keep an incident log recording start time, end time, cause and resolution for every outage
- Publish a public status page so customers and staff know what’s happening during an incident
- Test critical user journeys, such as checkout and form submissions, on a schedule rather than relying on a homepage check alone
- Instrument application logging so errors inside your CMS or database surface before customers report them
Guidance from GOV.UK’s service manual recommends designing for degraded operation, using cached pages or read-only modes so a partial failure doesn’t look like a total outage to visitors.
- Take timestamped screenshots and export monitoring logs whenever an incident occurs
- Store this evidence somewhere separate from the affected system, in case that system is what’s down
Providers rarely hand over service credits automatically, so the burden of proof usually sits with you, and a claim submitted without evidence is easy to reject.
Checklist: what to ask for and negotiate in an uptime SLA
Before you sign a hosting or maintenance contract, work through this list with whoever’s proposing the SLA.
- Agree the exact measurement method and how often you’ll receive reporting
- Confirm notification windows for planned maintenance and an emergency contact route for genuine incidents
- Get the claim process in writing, including deadlines and exactly how credits are calculated
- Ask for a defined Recovery Time Objective, alongside response and resolution targets for different severity levels
- Request audit rights so you can verify uptime figures independently rather than relying solely on the provider’s dashboard
Sample clauses in public-sector cloud contracts often allow suppliers to exclude maintenance from downtime calculations and require several days’ notice for scheduled work, so pin these details down before you sign rather than discovering them mid-outage.
MB Web perspective: practical resilience beyond the SLA
An uptime SLA reads like a safety net, but it’s really just a financial remedy after the fact. We think too many business owners treat the percentage on the contract as the whole plan, when it’s only the starting point.
- Proactive monitoring catches problems before customers do, rather than waiting for a probe to log a failure
- Routine maintenance and prompt patching close the gaps that cause outages the SLA never covers, like plugin conflicts or expired certificates
- A service credit might cover a fraction of a month’s hosting fee, but it does nothing for the sales or trust lost while your site was down
We’d rather help clients avoid the outage altogether than argue about the credit afterwards, which is why we treat monitoring and maintenance as part of the same job as hosting, not an optional extra bolted on later.
— James Golding
How MB Web can help with hosting, maintenance and monitoring
We manage website hosting with the maintenance and monitoring built in, rather than leaving you to chase a provider after something’s already gone wrong. Our hosting starts from a set yearly fee, with extra storage bolt-ons available for sites that outgrow their allowance; please visit our website for current pricing details.
Alongside hosting, our website maintenance and updates service keeps your CMS, plugins and dependencies patched, which closes many of the gaps an SLA never touches.
- We check the whole dependency chain, not just the server
- We flag risks before they become outages, using the kind of routine patching we cover in our piece on website maintenance and security
- We recommend realistic Recovery Time Objectives based on what your business can actually tolerate
If you’d like an honest look at where your current setup falls short, get in touch for a free website audit and we’ll talk you through the gaps and what closing them would involve.
Sources
- Gov
- MyNCSC help centre – Web Check – NCSC
- Hosting services agreement (contract attachment)
- RLI Application & Website Hosting – Digital Marketplace
FAQ
What does a 99.9% uptime SLA mean?
That works out to roughly 43 minutes of allowed downtime a month, though scheduled maintenance is usually excluded from that figure.
What is the standard SLA for uptime?
There’s no single universal standard, but minimum guaranteed uptime around 99.5% is common among hosting and cloud providers, with many premium services offering 99.9%. Always check whether the figure is measured monthly or annually, since that changes what it actually permits.
Is 99.5% uptime good?
For a site where every minute of downtime costs sales, such as an online shop, a tighter target like 99.9% is worth negotiating for.
How do I measure website uptime accurately?
Accurate measurement means monitoring from multiple external locations rather than trusting a single dashboard, and testing real user journeys like checkout or form submissions, not just the homepage. Keeping an independent incident log with timestamps also gives you evidence if you ever need to raise a service credit claim.