ℵ₀ CANTOR LABS
Company Products Developer Docs Changelog
ENTERPRISE SLA SPECIFICATION — CDX-SLA-2026-V1

Service Level Agreement

Enterprise Service Level Commitments & Availability Guarantees

Last Updated: September 7, 2026 | Effective Date: September 7, 2026
🏢 99.99% Sovereign Availability Guarantee:
Cantor Labs guarantees up to 99.99% monthly system uptime for Business and Enterprise Sovereign customers. In the event of a verified availability shortfall, eligible customers receive automatic contractual service credits.

1. Monthly Uptime Percentage Measurement

Monthly uptime percentage is calculated according to the following canonical formula:

Monthly Uptime (%) = [(Total Monthly Minutes - Outage Minutes) / Total Monthly Minutes] × 100
Plan Tier Target Uptime Max Permissible Downtime (Monthly) Support Response SLA
Developer Free 99.0% ~7.2 hours Community Forum Support
Team ($499/mo) 99.5% ~3.6 hours 24-Hour Email Response SLA
Business ($1,499/mo) 99.9% ~43.2 minutes 4-Hour Priority Response SLA (4h SLA)
Enterprise Sovereign 99.99% ~4.3 minutes 15-Minute Dedicated Hotline for P1 Incidents

2. Service Credit Schedule & Claim Rules

If monthly uptime falls below the target commitment, Customer is entitled to the following Service Credits applied against the subsequent billing cycle:

Actual Monthly Availability Service Credit Percentage
< 99.9% to ≥ 99.0% 10% of actual monthly fee
< 99.0% to ≥ 95.0% 25% of actual monthly fee
< 95.0% 50% (Maximum Monthly Cap) of actual monthly fee
⚠️ Mandatory Service Credit Claim Requirements:
1. Claim Window: Customer must submit a formal written claim with incident timestamps and telemetry logs to sla@cdxengine.com within thirty (30) calendar days following the end of the month in which the incident occurred. Failure to submit within 30 days constitutes a waiver of claim.
2. Offset Only: Service credits apply solely against future billing cycles and are non-refundable in cash and non-transferable.
3. Liability Cap: Cumulative service credits for any single month shall not exceed 50% of the monthly fee paid for that month.

3. Incident Severity Levels & Target Response Times

Severity Definition Initial Response Window
P1 (Critical) Complete production multi-agent cluster outage affecting critical business operations ≤ 15 minutes (Enterprise Sovereign) / ≤ 4 hours (Business)
P2 (Major) Significant degradation in payload slimming throughput or partial proxy unavailability ≤ 1 hour (Enterprise) / ≤ 8 hours (Business)
P3 (Moderate) Non-critical operational issues or minor feature anomalies with standard workaround ≤ 8 hours (Enterprise) / ≤ 24 hours (Team/Business)
P4 (Minor) General technical inquiries, documentation questions, or enhancement requests ≤ 24 hours

4. Exclusions from Downtime Calculations

Downtime excludes any unavailability resulting from:

  • Outages, rate limiting, or API deprecations by upstream third-party LLM providers (OpenAI, Anthropic, Google, etc.).
  • Customer self-hosted infrastructure crashes (e.g., vLLM or Ollama out-of-memory errors).
  • Customer private VPC, firewall, DNS, or network routing misconfigurations.
  • Scheduled maintenance announced at least 72 hours in advance.
  • Force Majeure events, internet backbone disruptions, or DDoS attacks.
ℵ₀ CANTOR LABS © 2026 Cantor Labs Inc. All rights reserved.
99.99% Operational
· Terms · Privacy