Service levels
Last updated 13 September 2026. This is a working draft written to be readable; it will be reviewed by counsel before general availability. If anything here is unclear, ask us and we will fix the wording rather than explain it away.
Availability
We target 99.5% monthly uptime, measured against successful responses to the application and API. That allows about 3.6 hours of downtime a month, which is a deliberately unambitious number for a product at this stage — we would rather publish one we can meet than one that reads well.
Excluded: planned maintenance announced at least 48 hours ahead and kept to under an hour, and outages caused by something upstream of us that we cannot route around.
Planned maintenance
Announced by email to account administrators, scheduled outside 06:00–20:00 US Central on weekdays wherever possible.
Support response
| Severity | Meaning | First response |
|---|---|---|
| 1 — down | Nobody can use the service | 1 hour, any day |
| 2 — blocked | A core workflow is unusable and has no workaround | 4 business hours |
| 3 — degraded | Something is wrong but there is a way round it | 1 business day |
| 4 — question | How do I, or a suggestion | 2 business days |
Business hours are 08:00–18:00 US Central, Monday to Friday. Severity 1 is answered outside them.
Backups and recovery
Full database and per-tenant backups nightly, checksummed, held seven days. Recovery point objective 24 hours; recovery time objective 4 hours for the platform, 1 hour for a single tenant restore.
If we miss it
Below 99.5% in a month, ask and we will credit 10% of that month's fee; below 99.0%, 25%. Credits are applied to the next invoice. We do not make you prove the outage — we have the same graphs you do.
Notification
Incidents affecting more than one customer are emailed to account administrators as they happen, with a written follow-up within five business days for anything over an hour.