Infrastructure and operations

We run dedicated servers and virtual machines hosted in Europe, under our direct administration. That choice guarantees data location, a predictable invoice, and no dependence on a provider whose pricing moves with usage.

Every machine is documented, its services are inventoried, its access is named. Infrastructure nobody knows how to rebuild is not infrastructure, it is a liability.

• Servers in Europe, under direct administration
• Inventory of services, ports and access
• Per-machine firewall and restricted administrative access
• Predictable monthly cost, no usage-based billing

A release should not be visible. We build each version in a fresh directory, verify it, then switch a symbolic link and reload processes one at a time. Requests in flight finish on the old version, the next ones land on the new one.

Rolling back follows the same path in reverse and takes seconds. That symmetry is what makes a calm Friday afternoon deployment possible.

• Offline build, atomic switch by symbolic link
• Progressive process reload, no request dropped
• Rollback to the previous version in seconds
• Previous releases kept and immediately reactivatable

A backup that has never been restored is not a backup, it is an assumption. We encrypt backups, copy them off the originating server, and periodically verify that a full restore completes within a known time.

The recovery time is written down and tested. It is the only figure that matters on the day you need it.

• Encrypted backups stored off the originating server
• Restore tested periodically, recovery time measured
• Retention defined by data category
• Recovery procedure written and executable by a third party

We watch availability, response codes, memory, disk space and certificates. Alerts that fire without anyone acting on them get removed: an alert people ignore destroys the value of every other alert.

Every incident produces a written note with its cause, its fix and the guard that prevents a repeat. That memory is an operations team's main asset.

• Availability and resource monitoring
• Automatic certificate renewal
• Written incident note, with cause and guardrail
• Security patches applied on a defined rhythm