1. Overview
This Service Level Agreement ("SLA") describes VoltageGPU's commitment to service availability and performance for our GPU cloud computing and AI inference services. This SLA applies to all paid customers with active subscriptions.
2. Service Availability
2.1 Uptime Commitment
GPU Pods (Compute)99.9% Monthly Uptime
API Services99.95% Monthly Uptime
Dashboard & Portal99.5% Monthly Uptime
2.2 Uptime Calculation
Monthly Uptime Percentage is calculated as:
Where "Downtime" means a period during which the service is unavailable, excluding scheduled maintenance and exclusions listed below.
3. Service Credits
3.1 Credit Schedule
If we fail to meet our uptime commitment, you may be eligible for service credits:
3.2 Credit Request Process
- Credits must be requested within 30 days of the incident
- Submit requests to support@voltagegpu.com
- Include affected service, date/time, and duration of outage
- Credits are applied to future invoices (not refunded)
- Maximum credit per month: 50% of monthly service fees
4. Exclusions
The following are not counted as downtime for SLA purposes:
- Scheduled Maintenance: Announced at least 48 hours in advance
- Emergency Maintenance: Critical security patches or urgent fixes
- Customer Actions: Issues caused by customer code, configuration, or misuse
- Third-Party Services: Outages of external services (payment processors, etc.)
- Force Majeure: Natural disasters, war, government actions
- Network Issues: Internet connectivity problems outside our infrastructure
- DDoS Attacks: Distributed denial of service attacks
- Beta/Preview Services: Features explicitly marked as beta or preview
5. Maintenance Windows
5.1 Scheduled Maintenance
- Standard Window: Tuesdays 02:00-06:00 UTC
- Notification: Minimum 48 hours advance notice via email
- Duration: Typically 1-2 hours, maximum 4 hours
- Frequency: Maximum once per week
5.2 Emergency Maintenance
- May occur at any time for critical security issues
- Notification provided as soon as reasonably possible
- Post-incident report within 48 hours
6. Support Response Times
Critical (Service Down)1 hour response
High (Major Impact)4 hours response
Medium (Partial Impact)8 hours response
Low (General Inquiry)24 hours response
Response times are measured during business hours (Monday-Friday, 09:00-18:00 CET), except for Critical issues which are monitored 24/7.
7. Performance Targets
7.1 API Performance
- Latency (P95): < 200ms for API endpoints
- Throughput: As specified in your plan limits
- Error Rate: < 0.1% for non-customer errors
7.2 GPU Pod Performance
- Provisioning Time: < 5 minutes for standard pods
- GPU Availability: Full GPU resources as specified
- Network: Minimum 1 Gbps connectivity
8. Monitoring & Reporting
- Status Page: Real-time service status at status.voltagegpu.com
- Incident Reports: Published within 48 hours of resolution
- Monthly Reports: Available upon request for enterprise customers
9. Customer Responsibilities
To benefit from this SLA, customers must:
- Maintain valid payment method and account in good standing
- Report issues promptly through official support channels
- Implement reasonable retry logic in applications
- Follow best practices for resource management
- Keep contact information up to date for notifications
10. SLA Modifications
VoltageGPU reserves the right to modify this SLA with 30 days notice. Changes will be communicated via email and posted on our website. Continued use of services after the effective date constitutes acceptance of the modified SLA.
12. Effective Date
Last Updated: January 2025
Effective: January 1, 2025