Service Level Agreement

Scope & definitions

This SLA applies to paid subscriptions of the Adfinia platform. Free tiers are best-effort and not covered.

"Service" means the Adfinia control plane (the web console, admin console, public marketing site, documentation) and the data plane (the API, dispatch workers, analytics ingest, AI gateway) covered by your subscription.

"Downtime" means a minute during which Customer-initiated requests to the data plane return a 5xx response or fail to connect for reasons within Adfinia's control, as measured by our external probes (§3). Downtime is counted in whole-minute increments per region.

"Monthly Uptime Percentage" means (total service minutes in the calendar month − Downtime minutes) ÷ total service minutes × 100.

Uptime targets by tier

The 99.5% "Demo window" target applies to pre-GA tenants on the AWS demo environment described in our deployment documentation. From the date of General Availability, every paid tenant moves to the 99.9% target by default.

For Enterprise customers requiring a 99.95% commitment, the elevated target is set out in the Order Form together with the corresponding credit cap.

How we measure

Uptime is measured by Adfinia's external probes hitting the public health endpoints in each region:

• GET /api/v1/health on the data plane.
• GET /healthz on the marketing and docs surfaces.
• A canary write-and-read transaction against a probe tenant on the API every 60 seconds.

A region is considered down only when at least two probes in independent network paths report failure for the same minute. Single-probe blips, transient connection resets, and packet loss attributable to upstream networks are excluded.

The status page at status.adfinia.com publishes the rolling 90-day uptime per region.

Service credits

If the Monthly Uptime Percentage for a region falls below the applicable target, the affected Customer is entitled to a service credit calculated against the fees paid for the affected region in the month of the incident:

Service credits are issued against future invoices, are not redeemable for cash, and do not extend the subscription term. Service credits are the Customer's sole and exclusive remedy for any failure of Adfinia to meet the SLA, except where mandatory law provides otherwise.

The maximum credit Adfinia will issue in a single month is capped at 50% of the monthly fee for the affected region.

Exclusions

Downtime does not include — and credits are not payable for — outages caused by:

• Planned maintenance announced under §6.
• Force-majeure events as defined in the Terms of Service.
• Failure of Customer-provided dependencies (the Customer's own ad accounts, payment processor, IdP, DNS, or a Customer-supplied LLM endpoint).
• Failure of upstream third-party networks, internet backbones, or cloud-provider zones outside Adfinia's control, provided we have followed our documented multi-AZ resilience pattern.
• Suspensions properly applied under the Terms of Service (non-payment, AUP violation, lawful order).
• Beta, preview, or "labs"-flagged features, which are explicitly outside the SLA until they leave preview.
• Configuration errors made by the Customer in the Platform (for example, deleting their own tenant, revoking their own credentials, or pinning to an unavailable region).

Planned maintenance

Adfinia performs routine maintenance — base-image rebuilds, security patching, schema migrations — primarily within rolling maintenance windows:

Where maintenance is expected to affect availability, we publish notice at least:

• 72 hours in advance for routine maintenance.
• 5 business days in advance for any maintenance that requires a brief service interruption longer than 15 minutes.
• As soon as practicable for emergency maintenance required to respond to a security incident, regulator instruction, or zero-day vulnerability.

Notices appear on the status page and are emailed to the operational contact on the Customer's account.

Incident severity & communications

Adfinia classifies incidents into four severities:

All severities are published on the status page. Sev-1 and Sev-2 also page the operational contact on the Customer's account by email and (where configured) webhook or SMS.

Within 5 business days of resolving a Sev-1 or Sev-2 incident, Adfinia publishes a post-incident review summarising the timeline, root cause, customer impact, and remediation steps.

How to claim credit

1. Email support@adfinia.com with subject "SLA credit request — <tenant ID> — <month>".
2. Include the tenant ID, the affected region, the date range of the incident(s), and a brief description of impact.
3. Submit within 30 days of the end of the calendar month in which the incident occurred. Requests later than 30 days are out of scope.

Adfinia will confirm receipt within 2 business days and respond with a credit decision within 10 business days. Approved credits appear on the next invoice after approval.

Changes to this SLA

Adfinia may update this SLA from time to time. Material changes — for example, a reduction in the uptime target or a reduction in the credit schedule — will not take effect until the Customer's next renewal, and require at least 30 days' notice. Older versions are archived alongside the Terms history.