A maintained and monitored database with tested recovery

Your application runs, so the database runs too. Nobody looks after it, until the disk fills up, a report that was done in a second last year suddenly takes 20 seconds, or someone asks whether Tuesday's data can still be brought back. To that last question usually nobody knows the answer, because nobody has ever tried. We take over the management of your database, including the answer to that question.

Why a database quietly goes downhill

A database rarely breaks in one go. It grows. Tables that started with ten thousand rows hold eight million after four years, and a query that never needed an index now does. The disk slowly fills up with log files. A version that was once current no longer gets security updates. None of that gives a notification at the moment it goes wrong.

We measure that before you notice it. We look at the size and the growth per month, at the slowest queries, and at the memory the database actually uses. Often the fix is small: an extra index, a setting still at the default value from ten years ago, or a table that can be cleaned up. Sometimes the conclusion is that the machine underneath has become too small, and then we tell you right away instead of working around it for six months.

Backups you can actually restore

A backup that has never been restored is an assumption. We make a full copy every night, keep it for a period we record together, and periodically restore it to a separate environment. There we measure how long it takes. That recovery time is on paper, so you know it on a quiet Tuesday and do not have to discover it at the moment things go wrong.

We store the copies in a different location than the database itself, because a backup next to the server does not help you if that server drops out. If you want the same arranged for your files, mailboxes, and virtual machines, that is Backup and disaster recovery; here it is specifically about the database and the data in it.

Updating without the business standing still

Updates are left undone because nobody wants to take the risk. That is understandable: a version jump of MySQL or PostgreSQL can break a query that has worked for years. So we first run the new version next to the old one, with a copy of your own data in it, and let your application talk to it. What breaks, breaks on the test environment.

The switch itself we plan in a window that suits you, with the old environment still on hand. Monitoring continues afterward: we watch the connections, the replication, the free space, and the duration of the heavy queries. If something goes wrong, an alert goes out according to an escalation we agree in advance, so it is clear who gets called and when.

Your data stays yours and stays in the EU

A database often holds the most sensitive things you have: customer data, orders, records, payroll lines. We run your database in our managed environments, and the data does not leave the EU. Access is limited to the people who really need it, connections are encrypted, and we log who did what.

You stay the owner. The schema, the data, and the scripts we set everything up with are yours, and you can get a full dump at any time to continue elsewhere. We do not sell a lock on the door. If you want to do the management yourself later, we hand it over including documentation of the settings and the recovery procedure.