Service Level Framework

In short: this is how we measure availability, how fast we answer by severity (a Severity 1 gets a first response within 4 business hours), how we announce maintenance and incidents, and how service credits work. The figures bind us only where your order form states them. We publish no marketing uptime number. In short boxes are a plain-language guide. If a summary and the full text differ, the full text applies.

We publish the framework, not a marketing uptime number. A percentage means nothing without a public status page and a measurement method, and we do not run one yet. The commitment that binds us is the one written in your order form.

1. Availability

In short: we measure whether the health endpoint answers, in five-minute intervals, each month. A target and its credits apply only where your order form states them. Announced maintenance, your own connected services, and your own network do not count against it.

Availability is measured monthly as the percentage of five-minute intervals in which the application's health endpoint responds successfully, excluding the exclusions below. Where an order form states a target, that target and its service credits apply. Without an order form commitment, the service is provided on a commercially reasonable efforts basis.

Excluded from the measurement: scheduled maintenance announced in advance; emergency maintenance needed for security; failures of a third-party service the customer connects, such as Microsoft 365 or an email provider; the customer's own network or configuration; force majeure; and suspension under the terms of service or the acceptable use policy.

2. Support targets

In short: a first response within 4 business hours for a critical outage, 1 business day for a serious impairment, 2 for a normal issue, 5 for a question. Business hours are 09:00 to 18:00 Central European Time, Monday to Friday.

How to read this table: each row is a severity level, what qualifies, and how fast a person first responds.

Severity	Definition	Target first response
Severity 1, critical	The service is unavailable or a core function is unusable for all users of the workspace, with no workaround	4 business hours
Severity 2, high	A core function is significantly impaired or unavailable for many users, with a difficult workaround	1 business day
Severity 3, normal	A function is impaired for some users, or a question that blocks work	2 business days
Severity 4, low	A question, a cosmetic issue, or a feature request	5 business days

Business hours are 09:00 to 18:00 Central European Time, Monday to Friday, excluding public holidays in Bulgaria and Germany. Enterprise order forms may set different hours, channels, and targets. Support is by email at the addresses on the contact page, and through the in-product feedback surfaces.

3. Maintenance

In short: routine releases happen without downtime where possible, and signed-in users see a notice before a restart. Anything that will interrupt service is announced at least 48 hours ahead. Security fixes can happen at any time.

Routine deployments are made without downtime where possible; signed-in users receive an in-application notice before a restart. Maintenance expected to interrupt service is announced at least 48 hours in advance to the account's administrative contact and scheduled outside business hours where practical. Emergency maintenance for a security issue may happen at any time, with notice as soon as reasonably possible.

4. Incident communication

In short: during a Severity 1 incident your administrative contact hears from us at least every two hours, and gets a written summary within five business days.

During a severity 1 incident we notify the administrative contact, give updates at least every two hours until service is restored, and provide a written summary within five business days covering what happened, the impact, the cause, and what changes as a result.

5. Service credits

In short: where your order form sets an availability target and we miss it, you get a credit on the next invoice, on the scale in that order form. Ask in writing within 30 days of the affected month.

Where an order form states an availability target, failing to meet it entitles the customer to a service credit against the next invoice, calculated on the tiered scale in that order form. Credits are the sole remedy for missed availability, are requested in writing within 30 days of the affected month, and are capped at the monthly fee for the affected service.

6. Backups and recovery

In short: the database is backed up continuously with point-in-time restore. Recovery objectives for enterprise customers are written in the order form. We do not publish a figure we have not tested for your deployment.

The managed database provider takes continuous backups with point-in-time restore. Recovery point and recovery time objectives for enterprise customers are stated in the order form. We do not publish a figure we have not tested end to end for a given deployment.

7. Dedicated instances

In short: a dedicated instance gets its own availability measurement, maintenance window, and release cadence, agreed in the order form and alignable to your change calendar.

A dedicated instance has its own availability measurement, maintenance window, and release cadence, agreed in the order form. Release cadence can be aligned to the customer's change calendar.