1. Scope
This SLA applies to the production Verus API and web application once deployed (AWS eu-west-2 by default, or the region selected in your order form). It does not apply to trial or sandbox environments, to the markst.com marketing surface, or to any third-party service you integrate with Verus.
2. Uptime commitment
Monthly Uptime Target: 99.5% on a calendar- month basis. This translates to a maximum ~3.6 hours of unscheduled production downtime per 30-day month.
Uptime is measured as the fraction of minutes in the month during which the service responded to an authenticated valuation request within 30 seconds with a 2xx response, observed from at least two independent geographic regions (UK and EU). A minute fails if it fails in both regions.
3. Maintenance windows
Planned maintenance windows do not count as downtime. We use them sparingly:
- Standard window. Sundays 02:00–06:00 UK time, up to once per calendar month. Announced in-product and by email to customer contacts at least 7 calendar days in advance.
- Emergency window. For a security patch or vendor-led incident response, we may initiate an unscheduled window. We notify customer contacts as soon as reasonably possible and publish a post-hoc explanation within five business days.
- Cap. Total maintenance downtime will not exceed 4 hours per calendar quarter.
4. Excluded events
Downtime caused by any of the following is excluded from the uptime calculation:
- Planned maintenance in a properly-announced window.
- Force majeure (natural disaster, war, large-scale internet outage, regulator-mandated shutdown).
- Customer-caused issues (invalid inputs, unauthorised API usage, customer-side network or IdP problems).
- Upstream-vendor outages beyond our control (AWS regional outage, Cloudflare edge outage, Anthropic API outage where AI features are enabled). We publish the impacted-vendor list in the incident post-mortem.
- Beta features clearly labelled as such.
5. Service credits
Where we fail to meet the 99.5% monthly target (excluded events excluded), customers receive a service credit against the monthly fee for the affected month:
| Monthly uptime | Service credit |
|---|---|
| 99.5% ≤ uptime | No credit due |
| 99.0% ≤ uptime < 99.5% | 10% of monthly fee |
| 95.0% ≤ uptime < 99.0% | 25% of monthly fee |
| uptime < 95.0% | 50% of monthly fee |
Credits are the sole and exclusive remedy for SLA breach, subject to the liability cap in the Terms of Service. Requests must be submitted within 30 days of the end of the affected month via support@markst.com and are applied against the next invoice.
6. Incident communication
- Detection & acknowledgement. Target acknowledgement within 15 minutes of an internally-detected incident; within 1 hour of a customer-reported one. Sev-1 incidents page on-call.
- Updates. Rolling status updates at least every 60 minutes during an active incident, via the status page and by email to affected-customer contacts.
- Post-mortem. A written post-mortem for any Sev-1 incident is shared within 5 business days, including timeline, root cause, and remediation. Available under NDA.
7. Backups & recoverability
Backup and recovery commitments sit in a separate statement at /legal/bcdr. In summary: RPO 1 hour, RTO 4 hours, multi-AZ within the primary region.
8. Review
This SLA is reviewed annually. Material changes are notified to customer contacts at least 30 days before taking effect. Enterprise customers may negotiate a higher uptime target (e.g. 99.9%) in the order form, subject to the corresponding architectural and commercial arrangements.
9. Contact
SLA questions or credit requests: support@markst.com. Commercial negotiation: legal@markst.com.