Service level agreement
The Nebbos service level agreement establishes uptime commitments, support response-time commitments, service credit remedies, and the scope of the commitments. Contract SLA text is derived from this reference; changes to the values below propagate to executed contracts at renewal.
Uptime commitments
| Tier | Monthly uptime | Maintenance |
|---|---|---|
| Team | 99.9% | Excludes scheduled maintenance windows announced at least seven days in advance. |
| Enterprise | 99.95% | Zero scheduled downtime. Maintenance operations execute as rolling restarts without impact to in-flight requests. |
Response-time commitments
| Severity | Description | Team | Enterprise |
|---|---|---|---|
| Severity 1 | Production unavailable | Acknowledgement within 4 business hours | Acknowledgement within 1 hour, 24/7 |
| Severity 2 | Major impairment | Acknowledgement within 1 business day | Acknowledgement within 4 hours, 24/7 |
| Severity 3 | Minor impairment | Acknowledgement within 2 business days | Acknowledgement within 1 business day |
| Severity 4 | Question or feature request | Acknowledgement within 5 business days | Acknowledgement within 2 business days |
Service credits
A monthly period in which actual uptime falls below the tier commitment generates a service credit against the subsequent month’s invoice.
| Measured uptime | Service credit |
|---|---|
| Below 99.9% and at or above 99.0% | 10% |
| Below 99.0% and at or above 95.0% | 25% |
| Below 95.0% | 50% |
Service credits apply to paid tiers only. Free-plan usage is provided without an SLA.
Exclusions
The following events are outside the scope of the SLA:
- Force-majeure events, including natural disaster, government action, and cross-provider outages affecting the underlying cloud region.
- Downtime attributable to operator-side misconfiguration, including revoked credentials, exhausted quota, and network routing errors.
- Capabilities explicitly designated as beta in the documentation or release notes. Beta features are opt-in and carry no SLA until they graduate to general availability.
- Cradle-side unavailability. The Cradle operates offline for reads; if the device is unavailable due to loss, damage, or replacement in transit, the platform’s uptime measurement continues to apply to the platform API rather than to the operator’s Cradle.
Uptime measurement methodology. Synthetic probes are executed against the platform API every 60 seconds from three geographic regions. A minute is recorded as unavailable if two of the three probes fail to return a 2xx or 3xx response within that minute. The measurement pipeline is publicly viewable at status.nebbos.ai . Every Severity 1 and Severity 2 incident is followed by a postmortem published within seven days of resolution.