Service Level Agreement

What we commit to

A monthly uptime target of 99.9% for the application and the API. That allows roughly 43 minutes of unplanned downtime a month.

Contractual service credits apply on Enterprise agreements. On Starter, Growth and Scale this target is a commitment we hold ourselves to and publish against, not a contractual guarantee. Presenting it as more than that would be dishonest.

What counts as downtime

The application or API returning errors, or failing to respond, for reasons within our control. Measured from our own monitoring, published on the status page.

• Not counted: scheduled maintenance announced at least 48 hours ahead.
• Not counted: outages at a third-party provider such as a marketplace or payment processor.
• Not counted: problems caused by your own configuration or by exceeding your rate limits.
• Not counted: force majeure.

Third-party dependencies

This is the honest caveat. CubieFlow depends on APIs we do not control. When Amazon throttles or Stripe has an incident, we queue, retry and tell you what is happening — but we cannot make their systems answer, and we will not pretend that outage is ours to fix.

What we do commit to is that nothing is lost. Queued work replays once the provider recovers.

Support response times

First response, during business hours in your region:

• Starter — 24 hours, email.
• Growth — 6 hours, email and chat.
• Scale — 2 hours, email, chat and phone.
• Enterprise — 1 hour, plus a named contact and an escalation path.

Incidents

Anything affecting multiple customers is posted to the status page within 15 minutes of detection, updated at least hourly, and followed by a public postmortem within five working days.

The postmortem describes what happened and what changed as a result. It does not name individuals, because incidents are caused by systems that allowed a mistake, not by the person who made it.