Service Level Agreement
Version 3.0. Published 7 September 2026. See the Policy Version History for applicability.
1. Covered availability
This SLA applies to the paid production service identified in your order. The following targets apply unless the accepted order expressly identifies a different service-specific availability commitment.
| Service | Monthly availability target |
|---|---|
| Shared, BDIX web hosting, Reseller, unmanaged VPS, Business Email | 99.9% |
| Managed VPS and Dedicated Servers | 99.95% |
| IP/proxy and n8n | 99.5% |
| AI-agent VPS | Underlying VPS target in its order; no guarantee of an external AI provider |
| Domains, external SaaS, development work | Only the expressly stated service-specific commitment |
The service boundary must be identified in the order: web-hosting platform availability; VPS power, virtualisation, assigned storage and network; dedicated-server contracted hardware/network; mail platform send/receive availability; or authenticated proxy connectivity. An operational ping alone does not establish availability where the covered platform is unusable. Customer application correctness and third-party destination availability are outside that boundary.
A covered outage starts when the covered service becomes unavailable and ends when service is restored. A transient interruption must last at least two consecutive minutes to qualify; once qualified, its entire duration counts. We use our records and reasonably reliable customer evidence. Our monitor is not conclusive in the face of substantiated contrary evidence.
2. Calculation and exclusions
Availability percentage = 100 × (M − E − D) / (M − E), where M is minutes in the paid covered portion of the calendar month, E is excluded minutes, and D is covered unavailability within the remaining minutes. Overlapping incidents and exclusions are not counted twice. If there are no eligible minutes, no SLA percentage is calculated for that period.
Exclusions are limited to downtime caused by:
- Routine notified maintenance, with at least 72 hours’ notice, within the disclosed window and up to four hours per calendar month under this SLA. Excess time counts unless a separate exclusion applies.
- Necessary urgent maintenance that could not reasonably await notice; only the reasonably necessary time is excluded, with an explanation afterward.
- Customer-controlled code, changes, unsupported configuration or usage exceeding an accepted plan limit, to the extent that caused the outage.
- Lawful suspension under the contract, or an external event meeting the Terms’ force-majeure test.
- A destination website, external mail recipient, customer-selected DNS/CDN, or external network outside our service boundary, to the extent responsible for the failure.
We do not automatically exclude every DDoS incident or upstream failure. Purchased mitigation, reasonable supplier selection and our own contribution must be considered. Overloading our own platform is not customer fault merely because demand increased.
3. Credits
First compare actual availability U with the applicable target T. If U meets or exceeds T, no credit is due. If U is below T, use the following complete bands without rounding up before comparison:
| Actual availability, where U < T | Credit on affected month’s service fee |
|---|---|
| U ≥ 99% | 10% |
| 98% ≤ U < 99% | 25% |
| 95% ≤ U < 98% | 50% |
| U < 95% | 100% |
Example: 99.93% availability fails a 99.95% target and gives 10% credit; 99.6% meets a 99.5% proxy target and gives no credit. Annual fees are allocated monthly. For partial months, use the actual covered period and its allocated fee. The total contractual credit is capped at that covered month’s affected-service fee and does not stack per incident.
Submit a claim through billing support within 30 calendar days after the end of the affected month, identifying the service and reasonably available incident evidence. This is an administrative deadline for this additional contractual credit, not a shortened deadline for a legal claim. We assess late requests reasonably where our own incident record establishes the outage.
Credits normally reduce the next invoice for that service. If we end the service before a promised credit can be used, we settle it against the final account or refund the balance rather than cancel it through our own termination. Unrelated account disputes do not automatically erase an accrued credit.
These credits are the agreed contractual remedy for ordinary covered downtime, subject to the Terms’ exceptions, express refund rights and all non-waivable legal remedies. A data-security breach or separate material failure is not automatically converted into an uptime-only claim.
4. Support and incident communications
Contact support@itnuthosting.com or open a priority ticket for an outage. The Service Scope and Backup Policy describes support channels and scope. First-response targets are not fixed resolution promises. We provide reasonable incident updates and, for a significant provider incident, an appropriate explanation once established; confidentiality and other customers’ security may limit detail. We do not promise an investigation will always finish within 72 hours.