Service Level Agreement
Uptime commitments, incident severity definitions, response and restoration targets, support hours, exclusions and the service credit schedule.
In short: 99.9% monthly availability on standard plans and 99.95% on enterprise, measured per tenant. Severity 1 gets an immediate 24/7 response. Miss the target and you get service credits, claimable within 30 days.
1. Where this applies
This Service Level Agreement ("SLA") applies only where you hold a paid subscription or managed services agreement with Nexiv Labs that expressly incorporates it. It does not apply to free trials, evaluation environments, beta features, sandbox tenants, or use of our public website.
It forms part of the Terms of Service. Where a signed order form states different figures, that document prevails.
2. Definitions
- Service — the production instance of the Nexiv platform or managed infrastructure covered by your agreement.
- Available — the Service accepts and correctly responds to valid requests.
- Downtime — any period in which the Service is not Available, excluding events under section 11.
- Monthly Uptime Percentage — (total minutes in the month minus Downtime minutes) ÷ total minutes in the month, expressed as a percentage.
- Incident — an unplanned interruption or reduction in quality of the Service.
- Response — a named engineer acknowledging the incident and beginning work, not an automated ticket receipt.
- Restoration — the Service returning to normal operation, whether by fix or by workaround.
- Business hours — 09:00–18:00 UK time, Monday to Friday, excluding England & Wales public holidays.
3. Availability commitment
| Plan | Monthly uptime | Max monthly downtime | Architecture |
|---|---|---|---|
| Standard | 99.9% | ≈ 43m 50s | Multi-zone, single region |
| Enterprise | 99.95% | ≈ 21m 55s | Multi-region active-active |
| Clinical | 99.95% | ≈ 21m 55s | Multi-region, region-pinned data |
| Trial / sandbox | No commitment | — | Best effort |
4. How we measure availability
Availability is measured per tenant, per calendar month, from our server-side telemetry: successful responses divided by total valid requests, sampled continuously and aggregated in one-minute buckets. A minute counts as Downtime if the error rate for that tenant exceeds 5% for the whole minute.
Synthetic checks run against core user journeys from multiple external locations at one-minute intervals as a secondary signal. Where our telemetry and your own monitoring disagree, we will review your data in good faith and explain any difference.
5. Performance objectives
| Objective | Target | Measurement |
|---|---|---|
| API latency | p95 < 300 ms | Server-side response time, excluding media streams and file transfer |
| API latency | p99 < 800 ms | As above |
| Page load | LCP < 2.5 s | Real user monitoring, 75th percentile |
| Media session setup | < 3 s | Consultation join to first frame |
| Change failure rate | < 5% | Deploys requiring rollback or hotfix |
Performance objectives are commitments of engineering practice and are reported monthly. They do not attract service credits — credits are tied to availability under section 12.
6. Incident severity
| Severity | Definition | Example |
|---|---|---|
| Sev 1 — Critical | Service unavailable or unusable for most users, confirmed data exposure, or a clinical safety impact. | Consultations cannot start; checkout is failing platform-wide. |
| Sev 2 — High | Major function impaired, or severe degradation, with no practical workaround. | Prescriptions cannot be issued; marketplace settlement is stalled. |
| Sev 3 — Medium | A function is impaired but a workaround exists, or performance is degraded without failing. | Search is slow; a report will not export in one format. |
| Sev 4 — Low | Minor defect, cosmetic issue, question or feature request. | A label is wrong; documentation is unclear. |
We assign severity on the facts. If you disagree with an assignment, say so — we will reassess immediately, and where it is genuinely arguable we take the higher severity.
7. Response and restoration targets
| Severity | Response | Update frequency | Restoration target | Coverage |
|---|---|---|---|---|
| Sev 1 | 15 minutes | Every 30 minutes | 4 hours | 24/7/365 |
| Sev 2 | 1 hour | Every 2 hours | 8 hours | 24/7/365 |
| Sev 3 | 4 business hours | Daily | 3 business days | Business hours |
| Sev 4 | 1 business day | On change | Next scheduled release | Business hours |
Restoration targets are objectives we work to, not guarantees — some root causes cannot be honestly bounded in advance. What we do guarantee is the response time, the update cadence, and that a named incident lead owns it until it is closed.
Every Sev 1 and Sev 2 incident receives a written post-incident review within 5 business days, sent to affected customers, covering timeline, root cause, customer impact and dated remedial actions with named owners.
8. Support hours and channels
| Channel | Availability | Use for |
|---|---|---|
| Email — support@nexivlabs.com | Monitored 24/7 for Sev 1 & 2 | All severities; primary channel |
| Phone — +1 (323) 289-0039 | Business hours | Urgent matters and escalation |
| Escalation contact | 24/7 (Enterprise & Clinical) | Sev 1 out of hours; details issued at onboarding |
| Status page | Always | Platform-wide incident visibility |
Mark Sev 1 emails "SEV1" in the subject line — that routes straight to the on-call engineer rather than the standard queue.
9. Planned maintenance
- Standard maintenance window: Sundays 02:00–06:00 UK time.
- At least 5 business days' notice for maintenance expected to cause any disruption.
- Most changes are deployed progressively with zero downtime and require no window at all.
- Notified maintenance inside the standard window does not count as Downtime.
- Emergency maintenance may be performed without notice where necessary to preserve security or data integrity; we notify as soon as practicable and it does not count as Downtime.
10. Backup, recovery and continuity
| Commitment | Target |
|---|---|
| Recovery Point Objective (RPO) | ≤ 5 minutes |
| Recovery Time Objective (RTO) | ≤ 60 minutes for region failover |
| Backup retention | 35 days point-in-time, in-region |
| Restore testing | Monthly automated, quarterly full exercise |
| DR exercise | Quarterly, results reported to Enterprise and Clinical customers |
Clinical tenant backups never leave the pinned region. A recovery plan that has not been executed is a document, not a plan — so we execute ours on a schedule and report what failed as well as what worked.
11. Exclusions
Downtime does not include unavailability caused by:
- notified planned maintenance, or emergency maintenance under section 9;
- your acts or omissions, including misconfiguration, invalid requests, or exceeding agreed capacity or rate limits;
- failure of your own systems, network or internet connectivity;
- third-party services you direct us to integrate with, where the fault lies with that third party;
- suspension properly applied under the Terms of Service or Acceptable Use Policy, including for non-payment;
- force majeure events beyond our reasonable control;
- trial, sandbox, beta or preview features, which are explicitly out of scope;
- attacks originating from or enabled by your credentials or infrastructure.
Where an underlying cloud provider fails, we still owe you our commitment: we designed for multi-region operation precisely so that a single provider region is not an excuse.
12. Service credits
| Monthly uptime achieved | Credit (Standard, 99.9%) | Credit (Enterprise / Clinical, 99.95%) |
|---|---|---|
| ≥ target | None | None |
| 99.0% – below target | 10% of monthly fee | 15% of monthly fee |
| 95.0% – 98.99% | 25% of monthly fee | 30% of monthly fee |
| 90.0% – 94.99% | 50% of monthly fee | 60% of monthly fee |
| Below 90.0% | 100% of monthly fee | 100% of monthly fee |
Credits are calculated on the monthly fee for the affected Service only, excluding pass-through infrastructure charges and one-off professional services fees. Credits are applied against your next invoice and are not paid in cash. Total credits in any month are capped at 100% of that month's fee.
Service credits are your sole and exclusive financial remedy for failure to meet the availability commitment. Where availability falls below 95% in three consecutive months, you may terminate the affected Service immediately without penalty and receive a pro-rata refund of prepaid fees.
13. Claiming a credit
- Email support@nexivlabs.com with "SLA CLAIM" in the subject line within 30 days of the end of the affected month.
- Include the affected Service, the dates and times, and any incident reference numbers.
- We validate against our telemetry and respond within 10 business days.
- Approved credits appear on the next invoice.
Where our own monitoring identifies a breach affecting you, we apply the credit without waiting for a claim. The claim process exists for cases we missed.
14. Reporting and review
- Monthly: availability and performance report per tenant, issued to your registered technical contacts.
- Quarterly: service review covering trends, capacity, planned changes and any outstanding remedial actions — for Enterprise and Clinical plans.
- Annually: review of this SLA against actual performance, with any proposed changes notified 30 days in advance.
- Ad hoc: raw SLO data for your tenant is available on request at any time.
15. Contact
- Support: support@nexivlabs.com
- Phone: +1 (323) 289-0039
- Status: System status
- Post: Nexiv Labs, Lees Street, Pendlebury, M27 6BU, GB