TrompTechLast updated: 29 September 2026
1. What this covers
These arrangements apply to the TrompTech Layer and to websites that TrompTech hosts for you, with a Hosting, All-in or Website per month subscription. They form part of the general terms and conditions. We only promise here what we actually do and can measure.
2. Availability
The Layer runs on a server in Germany. Each workspace has its own process, which restarts automatically after an outage. If the Layer is briefly not working, you see a maintenance page instead of an error message.
Our aim is that the Layer and your website are reachable at least 99.5% of the time per calendar month. That is at most about 3.5 hours of interruption per month. If we do not achieve that, you receive service credits (Article 7).
3. How we measure
Our uptime check requests your workspace every minute and your website every five minutes via the real address, as a visitor does: the name, the secure connection and the certificate. During an outage we measure every minute. At the second failed measurement in a row TrompTech receives an alert, and again when it works.
That measurement runs on our own server. If that server is completely down, nothing is measured. Time during which the measurement was not running is therefore counted as unavailable, unless we can show that the service was working at the time. On request we send you the overview for a month.
We post outages and planned maintenance on /status. We update that page by hand.
4. Maintenance
We preferably install updates outside office hours. Usually you do not notice them. Maintenance during which the service does not work for longer than 15 minutes is announced at least 2 working days in advance by email to the administrator of your workspace. Announced maintenance of up to 4 hours per month does not count towards availability.
5. Backups
Every night we make an encrypted backup of each workspace that leaves the server. We keep it for 60 days. The keys to open them are not held by the storage service. In addition, the server keeps track of every change, so that after an error we can go back up to 30 days.
In the event of an outage we restore the last good backup. If the entire server fails, anything entered after the last nightly backup may be lost. We test restoring at least twice a year.
6. Support and response times
You can reach TrompTech via support@tromptech.nl and, in the event of an outage, by telephone on +31 6 17720360. Office hours are Monday to Friday, 9.00 to 17.00, except on Dutch public holidays. The response times only apply within office hours.
| What | Example | First response |
|---|---|---|
| Outage | The Layer or your website does not work, or nobody can log in | Within 4 office hours |
| Disruption | A function does not work, but there is a workaround | Within 1 working day |
| Question | How something works, or a request | Within 2 working days |
A first response is not yet a solution. We then let you know what is going on and what we are doing. TrompTech is a sole proprietorship: during holidays we let you know in advance how outages will be handled.
7. Service credits
If a service is reachable less than 99.5% of the time in a calendar month, you receive a credit on the monthly fee for that service:
| Availability in the month | Credit |
|---|---|
| Below 99.5% | 10% of the monthly fee |
| Below 99.0% | 25% of the monthly fee |
| Below 97.0% | 50% of the monthly fee |
Request the credit within 30 days of the end of the month via support@tromptech.nl. We offset it against your next invoice. The credit is at most 50% of the monthly fee per month, and is the compensation for unavailability. For everything else, the arrangements on liability in the general terms and conditions apply.
8. What falls outside these arrangements
- Outages at services you connect yourself, such as Gmail, Microsoft 365, Mollie or a CRM.
- Outages at the AI provider. If it fails, the rest of the Layer keeps working.
- Outages in logging in via Supabase.
- Your own internet, devices or network.
- Changes that you or your team make to code, settings or the domain name.
- Announced maintenance, up to 4 hours per month.
- Force majeure, as in the general terms and conditions.
Even then we report what we see and help where we can.