This Service Level Agreement and Warranty defines the performance commitments, uptime targets, response times, support tiers, and warranty terms that Gradio Safety LTD provides to clients. It applies to all managed services, hosting, and ongoing operational engagements — and sets out the remedies available when service levels are not met.
1. Purpose and scope2. Definitions3. Service availability and uptime targets4. Incident severity and response times5. Support tiers and coverage hours6. Maintenance windows and scheduled downtime7. Monitoring and measurement8. Service credits and remedies9. Warranty10. Warranty exclusions11. Incident management process12. Communication and notification13. Client responsibilities14. Limitations and liability15. Exclusions from this SLA16. Termination and transition17. Compliance and review18. Contact us
This SLA & Warranty sets out the service level commitments that Gradio Safety LTD makes to clients who purchase managed services, hosting, or ongoing operational support. It defines what we guarantee, how we measure performance, what happens when we fall short, and the warranty coverage for deliverables.
This SLA applies to:
This SLA does not apply to one-time consultations, standalone architecture reviews, or stack recommendations unless explicitly stated in the engagement agreement.
| Term | Definition |
|---|---|
| Service | Any managed hosting, software, data, or operational service provided by Gradio Safety LTD under an engagement agreement |
| Uptime | The percentage of total minutes in a billing month that the Service is available and responsive, excluding scheduled maintenance and excluded downtime |
| Incident | Any event that causes or may cause degradation, unavailability, or malfunction of the Service |
| Critical incident | The Service is completely unavailable or a core function is non-operational |
| Major incident | A significant function is degraded or unavailable, but the Service remains partially operational |
| Minor incident | A non-critical function is affected, or there is a cosmetic issue that does not impact core operations |
| Response time | The time from when an incident is reported to when a Gradio Safety LTD engineer begins investigating |
| Resolution time | The time from when an incident is reported to when the Service is restored or a workaround is provided |
| Scheduled maintenance | Pre-planned downtime for updates, patches, or infrastructure changes, communicated at least 72 hours in advance |
| Service credit | A credit applied to the client's next invoice, calculated as a percentage of the monthly fee, when uptime targets are not met |
| Warranty period | The period after delivery during which Gradio Safety LTD will remediate defects in deliverables at no additional cost |
| Defect | A failure of a deliverable to conform to the specifications agreed in the engagement agreement |
Gradio Safety LTD commits to the following uptime targets for managed services:
| Service tier | Uptime target | Measurement period | Maximum allowed downtime / month |
|---|---|---|---|
| Premium managed hosting | 99.95% | Calendar month | ~21.9 minutes |
| Standard managed hosting | 99.9% | Calendar month | ~43.8 minutes |
| Data pipeline / processing | 99.9% | Calendar month | ~43.8 minutes |
| Monitoring / alerting service | 99.9% | Calendar month | ~43.8 minutes |
| Development / staging environment | 99.5% | Calendar month | ~3.65 hours |
Uptime is calculated as: (Total minutes in month - Excluded downtime minutes) / (Total minutes in month - Scheduled maintenance minutes) × 100
| Severity | Definition | Response time | Resolution target | Support coverage |
|---|---|---|---|---|
| P1 — Critical | Service completely unavailable; data loss risk; security breach | 15 minutes | 2 hours | 24/7 |
| P2 — Major | Core function degraded or unavailable; significant business impact | 30 minutes | 4 hours | 24/7 |
| P3 — Minor | Non-core function affected; limited business impact | 2 hours | 1 business day | Business hours (09:00–18:00 UK) |
| P4 — Low | Cosmetic issue, feature request, or general enquiry | 4 hours | 3 business days | Business hours (09:00–18:00 UK) |
Response time is measured from the moment an incident is reported via the designated support channel. Resolution targets are best-effort commitments — we work continuously on P1 and P2 incidents until resolved, including outside business hours.
| Tier | Coverage | Channels | Response (P1) | Included in |
|---|---|---|---|---|
| Basic | Business hours (09:00–18:00 UK, Mon–Fri) | 1 hour (during business hours) | Standard managed hosting, development environments | |
| Professional | Extended hours (08:00–22:00 UK, Mon–Sun) | Email + phone | 30 minutes | Standard managed hosting (production), data pipelines |
| Premium | 24/7/365 | Email + phone + dedicated Slack channel | 15 minutes | Premium managed hosting, mission-critical systems |
To keep systems secure and up to date, Gradio Safety LTD performs scheduled maintenance. Maintenance windows are:
| Window | Day / time (UK) | Typical duration | Notification |
|---|---|---|---|
| Primary maintenance | Sunday 02:00–05:00 | Up to 3 hours | 72 hours in advance |
| Secondary maintenance | Wednesday 02:00–04:00 | Up to 2 hours | 48 hours in advance |
| Emergency maintenance | As needed | Variable | As soon as practicable (minimum 2 hours) |
Gradio Safety LTD monitors all managed services using the following methods:
Uptime is measured by the synthetic monitoring system. If monitoring detects a failure, it is automatically classified as an incident and routed to the on-call engineer. Clients may also report incidents manually.
If Gradio Safety LTD fails to meet the uptime target for a managed service in a given billing month, the client is entitled to a service credit applied to the next invoice. Credits are calculated as follows:
| Monthly uptime | Service credit | Example (on £1,000/month service) |
|---|---|---|
| Below 99.9% but above 99.5% | 10% of monthly fee | £100 credit |
| Below 99.5% but above 99.0% | 25% of monthly fee | £250 credit |
| Below 99.0% but above 95.0% | 50% of monthly fee | £500 credit |
| Below 95.0% | 100% of monthly fee | £1,000 credit |
Service credits are the sole and exclusive remedy for uptime failures under this SLA, except where the failure is caused by Gradio Safety LTD's gross negligence or wilful misconduct.
Gradio Safety LTD provides a warranty on all deliverables produced during an engagement. The warranty covers defects in the deliverables that cause them to fail to conform to the specifications agreed in the engagement agreement.
| Deliverable type | Warranty period | What is covered |
|---|---|---|
| Architecture review report | 30 days from delivery | Corrections to factual errors, missing sections, or recommendations that do not align with the agreed scope |
| Security audit report | 30 days from delivery | Corrections to factual errors, false positives, or findings that do not align with the agreed scope |
| Custom software / code | 90 days from delivery | Bug fixes for defects that cause the software to fail to meet the agreed specifications |
| Infrastructure / deployment | 30 days from go-live | Configuration fixes for issues that cause the infrastructure to fail to meet the agreed specifications |
| Data pipeline / processing | 30 days from go-live | Fixes for data processing errors, pipeline failures, or output defects |
| Consultation recommendations | N/A (advisory only) | No warranty — consultations are advisory and provided "as is" |
The warranty does not cover:
| Exclusion | Reason |
|---|---|
| Changes made by the client or third parties to the deliverable after delivery | We cannot warrant modifications we did not make |
| Defects caused by client-supplied materials, data, or specifications | We work to the specification provided; incorrect input produces incorrect output |
| Defects caused by third-party services, APIs, or dependencies that change after delivery | We do not control third-party services |
| Defects caused by the client's environment, configuration, or infrastructure | We cannot warrant environments we do not manage |
| Issues arising from use of the deliverable outside the agreed scope or specifications | The warranty covers the agreed use case only |
| Cosmetic issues that do not affect functionality | Warranty covers functional defects, not aesthetic preferences |
| Force majeure events | Acts of God, natural disasters, war, pandemics, government actions |
| Normal wear and tear or expected degradation of performance over time | Systems require ongoing maintenance; the warranty is not a substitute for a support contract |
| Stage | Action | Timeline |
|---|---|---|
| 1. Detection | Incident detected by monitoring or reported by client | Real-time (monitoring) or upon client report |
| 2. Triage | On-call engineer assesses severity and begins investigation | Within response time for the severity level |
| 3. Communication | Client notified with initial assessment and estimated resolution time | Within 30 minutes of triage (P1/P2) or next business day (P3/P4) |
| 4. Investigation | Root cause identified; fix or workaround developed | Per resolution target for the severity level |
| 5. Resolution | Fix deployed or workaround applied; service restored | Per resolution target |
| 6. Verification | Engineer confirms service is operational; client confirms | Within 30 minutes of resolution |
| 7. Post-incident review | Root cause analysis document; preventive actions identified | Within 5 business days (P1/P2) or next monthly review (P3/P4) |
Gradio Safety LTD communicates with clients through the following channels during incidents:
| Channel | Used for | Tier |
|---|---|---|
| All incident notifications, status updates, post-incident reports | All tiers | |
| Phone | P1/P2 incidents — direct communication with on-call engineer | Professional, Premium |
| Dedicated Slack channel | Real-time updates during incidents, ongoing communication | Premium |
| Status page | Public or private status page showing current service status | All tiers (if applicable) |
Communication cadence during an active P1/P2 incident:
To ensure we can meet our SLA commitments, clients must:
This SLA is subject to the following limitations:
The following are excluded from uptime calculations and SLA commitments:
| Exclusion | Details |
|---|---|
| Scheduled maintenance | Pre-planned downtime communicated at least 72 hours in advance |
| Force majeure | Acts of God, natural disasters, war, terrorism, pandemics, government actions, internet-wide outages |
| Third-party outages | Failures of upstream providers (Vercel, Cloudflare, AWS, GCP, Azure, upstream banking & network providers) that are outside our control |
| Client-caused issues | Downtime caused by client actions — misconfiguration, unauthorised changes, exceeding resource limits |
| DNS / domain issues | Failures of client-managed DNS or domain registration |
| Network issues outside our infrastructure | ISP outages, BGP routing issues, CDN edge failures not caused by our configuration |
| Client-supplied code or content | Crashes or errors caused by code, content, or data provided by the client |
| Security incidents caused by client | Breaches resulting from client-shared credentials, weak passwords, or non-compliance with security recommendations |
If either party terminates the engagement agreement:
This SLA is reviewed every 6 months by the Security & Compliance Lead and approved by the founder. SLA performance is tracked monthly and reported to clients on request. Material changes to this SLA are communicated to affected clients at least 30 days before they take effect.
v2.0 2026-08-28 Full rewrite — 18 sections: uptime targets by tier, incident severity matrix, support tiers, maintenance windows, monitoring, service credit table, warranty by deliverable type, warranty exclusions, incident management process, communication channels, client responsibilities, liability cap, SLA exclusions, terminationv1.0 2026-08-27 Initial placeholder
To report an incident, request an SLA report, or ask questions about this SLA & Warranty: