Corefinity Service Level Agreement
Version 1.2. Last updated 16 November 2026. Effective date: 17 November 2026.
This SLA applies to CoreCluster and CoreStack online Orders governed by our Terms of Service, and to signed Orders only where expressly incorporated. Your existing signed agreement prevails for its Services. The SLA sets contractual remedies and does not remove other remedies preserved by the Terms or mandatory law.
1. Definitions and scope
1.1 Capitalised words have the meanings in the Terms. Monthly Fees means the pre-tax hosting Fees for the affected CoreCluster or standalone CoreStack Site in the relevant calendar month, including its hosting add-ons and staging. For a prepaid Annual Plan, the monthly equivalent is one twelfth of its actual annual payment, plus separately billed hosting options. Unrelated services and third-party costs are excluded.
1.2 Downtime means minutes when a fault in infrastructure or software we manage prevents a production Site from responding normally or materially prevents its transactional function. It includes a failure affecting one or more production Sites on a CoreCluster; the whole cluster need not fail. Slow responses alone count only where the fault materially prevents normal transactional use. Application faults outside our managed responsibility follow section 4.
1.3 Monthly Availability = (all minutes in the calendar month minus Downtime minutes) ÷ all minutes in that month × 100. Downtime excludes only the minutes demonstrably attributable to section 4. For a service running for part of a month, use its running minutes. Overlapping incidents are counted once for each affected service.
1.4 Business Hours: 09:00 to 17:30 Europe/London time, Monday to Friday, excluding public holidays in England. An hour within that schedule is a Business Hour. Critical support operates continuously and is not limited to this schedule.
2. Availability commitment
2.1 We commit to 99.99% Monthly Availability for each CoreCluster and standalone CoreStack Site. A CoreStack Site hosted within your CoreCluster is measured as part of that cluster, avoiding a duplicate credit on the same Fees. The commitment is to the managed production hosting service.
2.2 The Control Panel, API and MCP Server are supported services; this SLA does not give them a separate availability percentage. Their failure counts as hosting Downtime where it causes a production failure under 1.2. Other failures remain subject to the support deadlines, reasonable-care duties and other contractual remedies. Staging and development do not have a separate production availability target.
2.3 This is not a promise of zero downtime or multi-zone failover. Every minute counts in the availability calculation; credits have the separate complete-15-minute calculation below. In a 30-day month 99.99% permits 4.32 minutes of Downtime, so a miss shorter than 15 minutes can breach the target without generating a credit. Other remedies are not excluded.
3. Measurement and reporting
3.1 We monitor production hosting with automated checks and alerts, and investigate customer reports. Incident start is the earliest time established by our monitoring, a support ticket or other reliable evidence; it is not delayed until the customer opens a ticket. The incident ends when normal service is restored. Evidence can establish an earlier start than a monitoring alert. We keep the incident records used for the calculation for at least the period in which a credit can be claimed and reviewed.
3.2 We provide the incident record, applicable exclusions and monthly calculation on request. You may challenge it with reasonably available evidence. Our records are not conclusive merely because we hold them. Open an Emergency ticket promptly so we can act; a late ticket does not erase demonstrated Downtime.
3.3 Simultaneous failures of several Sites in the same CoreCluster count as one period for that cluster. Separate affected clusters or standalone Sites are assessed separately. Fees refunded under the guarantee are not credited a second time.
4. Exclusions and maintenance
4.1 We exclude only the part of an incident caused by:
- notified planned maintenance satisfying 4.2, or necessary emergency maintenance satisfying 4.3;
- your application, changes, credentials or instructions outside our managed responsibility, including your failure to address a notified critical application vulnerability, to the extent these actually cause the outage;
- exhaustion of the capacity you purchased, demonstrated by the incident evidence, rather than merely reaching an ordering warning threshold;
- your own DNS, CDN, connectivity or third-party service outside our supply chain;
- a lawful and justified suspension under the Terms; or
- an event outside our reasonable control meeting the force-majeure test in the Terms, despite reasonable precautions and mitigation.
4.2 We give at least seven days' notice of routine/high-priority planned maintenance and at least 24 hours' notice of critical planned maintenance. We do planned work in the maintenance window: by default 00:00 to 04:00 UK time, or, where the Control Panel lets you choose another four-hour window, the one you choose, shown with its time zone. We minimise interruption. The exclusion covers the actual notified work and interruption, not the entire unused window or unrelated failures.
4.3 Emergency work without the normal notice is excluded only to the extent urgently necessary to address an immediate security, safety or stability threat. We notify you as soon as practicable, explain the need and minimise disruption. Routine overdue patching does not become exempt merely by being labelled an emergency.
4.4 Ordinary Google Cloud infrastructure failures count where they prevent our hosting service. There is no blanket exemption for every cloud, zone or Region outage. If a particular event meets 4.1's force-majeure test, we explain and evidence that causal exclusion. We remain responsible for faults in our architecture, configuration or maintenance and for reasonable resilience measures we promised.
4.5 The 90% threshold for refusing creation of new Sites or Environments does not itself make existing Sites unavailable or disqualify them from this SLA. Exclusions do not remove data-protection duties, our obligation to use reasonable care, or liability that cannot lawfully be excluded.
5. Service credits
5.1 If Monthly Availability is below 99.99%, we credit 10% of the affected Monthly Fees for each complete 15 minutes of qualifying Downtime, up to 100% of those Monthly Fees. The complete-15-minute units are calculated on total qualifying Downtime in the month, without double counting. We do not round uptime losses down for the availability calculation.
5.2 Example: 45 minutes of qualifying Downtime in a month produces a 30% credit. Ten minutes misses the availability commitment in a full month but produces no complete 15-minute unit and therefore no credit. This example does not restrict other remedies.
5.3 Claim by ticket, identifying the service, date and incident and any evidence you have. Please claim within 30 days after the month ends so we can process it promptly; this is not a forfeiture of statutory rights or of a substantiated contractual credit solely because that administrative period passed. We do not require logs only we can access.
5.4 We explain the outcome and apply an agreed credit to the next invoice, or correct the next available invoice if investigation takes longer. Total credits cannot exceed the affected Monthly Fees, and there is no duplicate credit for the same Fees and incident. A genuinely disputed invoice does not automatically forfeit credits.
5.5 Credits are not transferable. If the service ends before an earned credit can be used, we apply it against outstanding undisputed charges and refund the remainder to the original payment method within 14 calendar days. This closes the online account without forfeiting an earned remedy.
5.6 During a free Trial the hosting Fees are zero, so the monetary credit is zero. A refund under the guarantee is taken into account only to prevent double recovery of the same Fees. The availability and response commitments still describe our service; free status does not remove mandatory duties.
5.7 Credits are not your sole remedy. Sections 15 and 21 of the Terms and section 6 below preserve termination, refunds and other lawful claims. Compensation and credits for the same loss are not recovered twice. Failure to meet a support response deadline does not create a separate automatic service credit under this scale, but remains a contractual breach where the deadline applies.
6. Repeated serious failure
6.1 Following the existing hosting templates, if qualifying Downtime exceeds 24 hours in each of two consecutive calendar months, you may terminate the affected Services immediately by notice. Unused prepaid Fees and earned credits are refunded. No early-termination charge applies to this exit.
6.2 There is no new 60-day forfeiture deadline for that right. Give notice promptly after identifying the repeated failure. This clause adds to other material-breach remedies and does not require you to endure two months before exercising another right.
7. Support priorities and response deadlines
7.1 Tickets are available through the Control Panel at all times. Use Emergency for critical incidents; it pages our on-call engineers. If the Control Panel is unavailable, email support@corefinity.com or call +44 800 464 7255 (UK) or +1 833 818 8090 (US); an outage of the Control Panel itself alerts our on-call engineers automatically. We record when your report reaches us by any of these routes; a failure of our Control Panel cannot stop the response clock.
| Priority | Definition | First human response deadline |
|---|---|---|
| Critical / downtime | All or part of a production Site cannot transact, or a material performance failure may affect conversion; platform incidents threatening an equivalent loss of service are treated urgently | 30 minutes, 24 hours a day, every day |
| High | A smaller issue potentially affecting conversion without production downtime | One Business Hour |
| Medium / low | Other platform faults, questions or requests without those effects | Four Business Hours |
7.2 A response is an appropriately qualified human acknowledgement and initial assessment or request for information needed to investigate; an automatic receipt is not a response. These are response deadlines, not guarantees of resolution within those periods. For high and routine issues received outside Business Hours the clock starts when those hours reopen.
7.3 We explain any priority reassessment using the actual impact. We do not downgrade a critical incident merely to avoid the deadline. A senior engineer coordinates investigation where appropriate, and we keep you informed of material progress and the restoration outcome. Your prompt cooperation helps resolution but does not excuse a missed initial response.
7.4 Platform support is included in every online CoreCluster and CoreStack plan at no extra charge. It covers our hosting platform, its infrastructure and software stack, deployments and our management tools, including investigation to identify the cause. Work on your application code, application development and third-party remediation are paid work that we agree with you, with its price, before starting, unless already included in your Order. Existing signed enhanced service schedules keep their own broader scope.
8. Recovery and security
8.1 Backups, restores, responsibilities and incident notice are governed by Terms section 14 and the DPA. No separate recovery-point or recovery-time guarantee is created by this SLA. Security incidents are notified promptly and without undue delay under the DPA, independently of ticket priority. PCI acquirer reporting times are not a customer breach-notification deadline.
9. Changes and disputes
9.1 Material changes follow Terms section 25 and cannot retrospectively remove an earned credit. Disputes use Terms section 26; governing law, signed-agreement priority, liability and mandatory protections follow the Terms. Nothing here binds an authority or a data subject to a contractual credit remedy.