Server Management and DevOps

Server management is not setting up a machine once and walking away. It is daily attention to everything from operating system updates and a disk about to fill up to an expiring SSL certificate or a service that quietly stopped. We start either by provisioning a new server from a provider such as Hetzner or by taking over your existing one, and build a tidy, documented setup for your website, API, database and background services.

Services run in Docker containers, with each one's version and configuration recorded. That makes moving an application to a new server or rolling back a bad update a matter of minutes. In front of the services we put nginx as a reverse proxy with automatically renewing SSL certificates, and on Cloudflare we configure your domain, DNS, CDN caching and DDoS protection.

A new Linux server on the internet starts receiving automated scans within minutes. Our first steps are to disable password SSH logins in favour of keys, block direct root login, set up a firewall that leaves only the required ports open and ban repeated failed login attempts with fail2ban. Security updates are applied automatically, while major version upgrades are done in planned maintenance windows.

Databases and caches are closed off from the outside world and run on an internal network that only the application can reach. Each service runs under its own user with no more privileges than it needs, and passwords and keys never sit in plain text in config files. Server access is logged, and the permission setup lets you remove a departing team member's access in a single step.

A backup is only useful if it can be restored. Databases and files are backed up automatically, encrypted, and sent to a location independent of the server, with retention set to fit the project. The disaster recovery plan describes in writing the steps to bring the system back if the server is lost entirely, and restores are actually tested on a staging environment at regular intervals.

We check the availability of your site and API from the outside around the clock, and track CPU, memory and disk usage, response times and error rates. Logs are collected in one place, so finding the cause of a problem does not mean logging into server after server. When a disk is filling up, a certificate is close to expiry or a service stops responding, an alert reaches our team before it turns into an outage.

Part of our DevOps work is how your software reaches the server. Instead of updates done by copying files by hand, we set up a CI/CD pipeline that runs your tests when code is pushed, builds the Docker image and deploys the new version without downtime. Staging and production are kept separate and every release can be rolled back, so shipping an update stops being a risky event.

We begin by reviewing your current infrastructure and giving you a written report of gaps and risks. Fixes are then applied in order of priority without disrupting the live system. Server credentials and all configuration belong to you and are handed over with documentation describing what was set up and how it is managed. After that, monitoring, updates and incident response stay with us under a monthly management plan.