Service Level Agreement

This Service Level Agreement ("SLA") describes the availability, performance, and incident response commitments for the DecisionHost platform operated by DecisionLedger AI. This SLA applies to customers on Professional and Enterprise subscription plans.

1. Scope

This SLA covers the DecisionHost web application, REST API, and all platform services accessed through decisionworksmodels.com and associated subdomains. It does not cover third-party integrations, customer-configured webhooks, or features explicitly designated as "Beta" or "Preview."

Customers on the Starter plan receive best-effort availability but are not eligible for service credits under this SLA.

2. Availability

2.1 Uptime Commitment

DecisionLedger AI commits to 99.9% monthly uptime for the DecisionHost platform, measured as the percentage of total minutes in a calendar month during which the Service is available and responsive.

Monthly Uptime Percentage is calculated as:

(Total Minutes − Downtime Minutes) / Total Minutes × 100

2.2 Downtime Definition

"Downtime" means any period of five or more consecutive minutes during which the Service returns HTTP 5xx errors for more than 50% of requests, or is entirely unreachable. Downtime is measured by our external monitoring systems.

2.3 Exclusions

The following are excluded from downtime calculations:

• Scheduled maintenance: Maintenance windows announced at least 72 hours in advance via email and in-app notification, limited to 4 hours per month
• Force majeure: Events beyond our reasonable control, including natural disasters, war, government actions, and widespread internet outages
• Customer-caused issues: Downtime caused by customer actions, including misconfigured integrations, exceeding rate limits, or account suspension due to Terms of Service violations
• Third-party failures: Outages of third-party services (AWS regional failures, Stripe, Cognito) that are beyond our control, provided we take reasonable steps to mitigate impact

3. API Performance

3.1 Latency Targets

For standard API endpoints (authentication, CRUD operations, queries), DecisionLedger AI targets the following response times:

• P50 (median): < 500 milliseconds
• P95: < 2 seconds
• P99: < 5 seconds

Model execution endpoints (plugin runs, Monte Carlo simulations, optimization solvers) may have longer response times depending on input complexity and are excluded from these latency targets.

3.2 Rate Limits

API rate limits vary by plan tier and are documented in the API reference. Requests exceeding rate limits receive HTTP 429 responses and are not counted toward availability or latency metrics.

4. Incident Response

4.1 Severity Levels

• Critical (P1): Service is entirely unavailable or data integrity is at risk
• High (P2): Major feature is unavailable or significantly degraded for multiple users
• Medium (P3): Feature is degraded but workaround exists
• Low (P4): Minor issue with no significant impact on operations

4.2 Response Targets

• Automated detection: < 5 minutes for all severity levels (CloudWatch alarms + SNS alerting)
• P1 acknowledgment: < 15 minutes during business hours, < 30 minutes outside business hours
• P2 acknowledgment: < 1 hour during business hours
• P3 acknowledgment: < 4 business hours
• P4 acknowledgment: < 1 business day

4.3 Communication

During incidents, we provide status updates through our status page and email notifications to affected customers. Post-incident, we publish root-cause analysis reports for P1 and P2 incidents within 5 business days.

5. Service Credits

5.1 Credit Schedule

If monthly uptime falls below 99.9%, eligible customers may request service credits according to the following schedule:

• 99.0% – 99.9%: 10% credit of that month's subscription fee
• 95.0% – 99.0%: 25% credit of that month's subscription fee
• Below 95.0%: 30% credit of that month's subscription fee

5.2 Credit Request Process

To request a service credit, submit a request to support@decisionledgerai.com within 30 days of the month in which the downtime occurred. Include the dates, times, and description of the unavailability. We will review and respond within 10 business days.

5.3 Credit Limitations

Service credits are applied to future invoices and are not redeemable for cash. Credits may not exceed 30% of the monthly subscription fee for any single month. Credits are the sole and exclusive remedy for failure to meet uptime commitments.

6. Data Durability

Customer data is stored on AWS infrastructure with built-in redundancy. Database backups are performed daily with point-in-time recovery available for the prior 7 days (Professional) or 35 days (Enterprise). Audit logs are stored with S3 Object Lock for tamper-proof retention.

7. SLA Modifications

We may update this SLA with at least 30 days' written notice. Changes will not reduce commitments for the remainder of any active annual subscription term. The current version is always available at this URL.

8. Contact

For SLA inquiries or to request service credits, contact us at support@decisionledgerai.com.