Our commitment toreliability.
Availability, support, and performance expectations are documented by plan and deployment. We set commitments in customer agreements instead of publishing unsupported universal numbers.
Last updated: January 2026
Overview
This page describes the service practices and commercial terms that may apply to customers of our platform. Specific commitments vary by plan, deployment model, and customer agreement.
Availability and performance expectations are defined with reference to the customer's environment, workload, and support plan rather than a single public percentage or latency target.
Availability terms by plan.
Different tiers can carry different commitments depending on deployment and agreement structure.
Starter
Shared-platform availability commitments
Documented in your plan and support terms
Scale
Enhanced availability planning with priority support
Structured maintenance and escalation processes
Production
Enterprise commitments defined per deployment
Availability expectations documented in enterprise agreements
Performance expectations by workload.
Performance depends on workload shape, model path, and deployment architecture.
API Response Time
InteractiveOptimized for fast API interactions
Actual timing varies by workload, model, and deployment
Streaming Latency
Streaming-FirstBuilt for prompt time-to-first-response experiences
Depends on request shape, model path, and deployment
Throughput
ScalableCapacity planning for production traffic
Scaled based on plan, workload, and deployment architecture
Processing Time
Workload-DependentComplex AI operations vary by task and model path
Customer-specific expectations are set during onboarding
Help when you need it.
Response times based on issue severity.
| Priority | Description | Response | Resolution | Availability |
|---|---|---|---|---|
| P1 - Critical | Complete service outage or major functionality unavailable | Rapid escalation | Prioritized handling | Continuous coverage path |
| P2 - High | Major feature degraded, significant business impact | Priority response | Expedited handling | Priority support path |
| P3 - Medium | Feature partially impaired, workaround available | Business-day response | Standard handling | Business hours |
| P4 - Low | General questions, minor issues, feature requests | Standard response | Best effort | Business hours |
Planned downtime.
How we handle necessary maintenance windows.
Scheduled Maintenance
Advance Notice
Advance notice when practical
Maintenance Window
Sundays 2:00-6:00 AM UTC
Frequency
Monthly (if needed)
Emergency Maintenance
Advance Notice
As soon as possible
Maintenance Window
As required
Frequency
Only for critical security/stability
Agreement-specific remedies.
Remedies and credits, if offered, are defined in customer agreements rather than a single public schedule.
How to Review Terms
- 1Review your order form, support plan, or enterprise agreement
- 2Confirm the availability, support, and escalation terms attached to your deployment
- 3Contact your account team if you need custom enterprise commitments documented
What's not covered.
The following situations are excluded from SLA calculations.
- Scheduled maintenance communicated in advance when practical
- Emergency maintenance for security or stability
- Actions or inactions of the customer or third parties
- Internet connectivity issues outside our network
- Force majeure events (natural disasters, war, etc.)
- Beta features or pre-release functionality
- Customer-caused outages (API abuse, misconfiguration)
Incident Communication
During incidents, we communicate through multiple channels to keep you informed:
- Status Page: Real-time updates at status.disruptiverain.com
- Email Notifications: Alerts to account administrators
- Post-Incident Reports: Detailed analysis within 5 business days
Questions about service commitments?
Contact us to discuss plan-specific availability, performance expectations, and enterprise support commitments.