Skip to content
Book a demo
Menu

Legal · service commitments

There is no uptime percentage, and here is why.

A service level agreement is a promise with a credit behind it. Superiora Planner is in pilot on a single host that is deliberately switched off at night to keep the cost of a pilot honest, so a 99.9 % figure would be a number we could not mean. This page says what we do commit to instead, and what has to change before a real SLA can be offered.

Last updated · 2026-09-18 Pilot · no uptime SLA Blanks marked for owner review

01
Pilot hours

The platform is available during pilot hours.

07:00 to 19:00, Monday to Saturday, America/Winnipeg. Outside those hours the host is stopped: the application and this website do not answer, and neither does the sign-in page. It is not an incident and it is not a fault — it is a scheduled window, and the daylight-saving shift moves with the timezone rather than against it.

Nothing is lost in the window. The model’s event log is on disk, the state is rebuilt on start, and work in progress is where you left it.

Inside the window we aim for continuous availability and will tell you when we cannot. Status carries the current picture and the record of anything that went wrong.

Pilot hours are a cost decision, not a technical limit. A customer who needs the platform outside them — a team in another timezone, a month-end that runs into the night — should say so: it is a deployment change, not a rebuild. [Owner: whether extended or 24/7 hours are offered, and on what terms]

02
What we do commit to

Four commitments we can keep today.

Support response
A first human response within one working day of a support mail, during pilot hours. A problem that blocks planning work is picked up the same day. Support says what to put in the report so the first reply can be useful.
Backup and restore
The data volume is backed up nightly to object storage in the same region. A model’s state is an append-only event log plus snapshots, so a restore replays to the same state rather than to an approximation of it. Restore is on request during the pilot; target recovery point [Owner: RPO commitment], target recovery time [Owner: RTO commitment].
Notice of planned change
A release that requires a restart, a migration or a change to how something behaves is announced to the administrators on each account beforehand. What shipped is published on the changelog.
Your data out, on demand
A model leaves whole — Excel, CSV, and a query interface BI tools read — at any time, without asking us. If you would rather we produced the export, we will, within [Owner: export request turnaround].

03
What we do not commit to

Stated plainly, so nobody has to discover it later.

  • No uptime percentage and no service credits. Not during the pilot.
  • No 24/7 on-call. A page at 03:00 reaches nobody. Support is people, and the people are in São Paulo · London · remote.
  • No multi-region failover, and no hot standby. One host, one region, restored from backup.
  • No formal severity matrix. We triage by what is blocking your planning work, and tell you what we are doing.
  • No third-party certification. No SOC 2, no ISO 27001, no external penetration test report. Security says what is in place instead.

04
What changes it

The conditions for a real SLA.

A committed uptime number needs three things that a pilot does not have: a host that stays on, a standby to fail over to, and enough measured history to know what we are promising. When a customer needs those, they are a deployment and a commercial conversation rather than a product gap — say so on the order form and it becomes part of it.

Until then, the commitments on this page are the ones in force, and a signed order form may set different terms for your account.