Where’s the uptime percentage?
We’re deliberately not publishing an uptime SLA percentage until we have long-baseline, multi-vantage data to back one — most hosts’ 99.9% badges are marketing, not measurements.
Look closely at a typical hosting SLA and you’ll find it isn’t a promise your site stays up — it’s a schedule of small account credits, claimable only if you notice the outage, file within the window, and the host’s own measurement agrees with yours. A 10% credit on a £20 plan is £2 for the worst day of your quarter.
We think the honest version of reliability is: monitoring you can see, incidents we admit to in public, backups that prove themselves nightly, an architecture that keeps your site serving even when our own management plane fails — and your money back in the first 30 days if we don’t live up to it.
Five things we actually commit to
Each one is live on the platform today — not an aspiration.
1. Independent per-minute monitoring
Every hosted site is health-checked every minute by monitoring that runs independently of the site itself. Failures alert us — and you, by email, webhook or Slack — the minute they happen, with a daily digest confirming the quiet days. Brief outages are recorded, not averaged into invisibility.
How monitoring works2. A transparent public status page
Our status page fetches live from the platform API when you load it — it is not a static wall of green bars. Incidents appear on it while they are happening, and we do not scrub bad days from the record afterwards.
View live status3. Backups restore-tested every night
Every backup is automatically restored into a sandbox each night and verified. Not a checksum, not a file-count — an actual restore. When you need a backup, you are restoring from one that provably restores. We believe this should be the industry standard; it mostly is not.
How backup testing works4. Your site survives our control-plane outage
Routing on the worker machines that serve your site is local. If our management plane — the dashboard, the API, the orchestrator — goes down, customer sites keep serving traffic. We designed the architecture around our own worst day, and databases run hot-standby replication between workers as well.
5. 30-day money-back guarantee
Every paid plan carries a 30-day money-back guarantee. If the platform does not live up to this page in your first month, you get your money back — a commitment with actual money behind it, unlike a symbolic 10% service credit.
See plans & pricingWhen something breaks
The honest version of an incident process — what actually happens, without invented stopwatch targets.
Detection
Per-minute health checks fail and our alerting fires. We are usually investigating before the first customer email arrives.
Containment
Site-serving infrastructure is isolated from the management plane, so a control-plane problem does not become a your-site problem.
Transparency
The public status page reflects the incident while it is live. If your site is affected, your own monitoring alerts have already told you.
Fix & learn
We fix the actual system — it is ours, end to end — and document what happened internally so the same failure does not repeat.
Service commitment FAQ
The legal detail lives in our terms of service. Live platform state is on the status page.