Configuring Website Deployment on VDS/VPS

A developer spent a month fighting Laravel deployment on a VDS — every time something went wrong: either permissions on storage were reset, migrations didn't run, or npm ci failed with an error. The project stalled, and the client got nervous. Sound familiar? We solved it in two days by automating deployment and writing simple documentation. Full control over the server is great, but only when everything is set up properly. Our experience: over 6 years of server configuration and 300+ successful projects. We guarantee a 99.9% uptime and reduce deployment time by 90%.

What problems does professional VDS/VPS deployment solve?

Insecure root access. Many beginners leave SSH open for root with a password. Brute force tries thousands of combinations in a day. We create a separate deploy user with keys, disable root login, and set up Fail2ban — the risk of break-ins drops by 99%.

Manual deployments with errors. When every update means manually running git pull, composer install, npm run build, and hoping nothing breaks, a fatal error eventually happens. We set up automatic deployment via Deployer or GitHub Actions: one command pushes code, and the server automatically pulls updates, runs migrations, and reloads PHP-FPM. This approach is 10 times faster than manual: instead of 30 minutes, deployment takes 3.

Lack of monitoring. You learn about a site outage from the client, not from the system. We install Netdata or Prometheus Node Exporter — you see CPU load, memory, disk, and network in real time. Outages won't catch you off guard.

How we set up deployment: a case from our practice

Consider a project from our practice: the client rented a VDS from Selectel (4 vCPU, 8 GB RAM) for a Laravel app with PostgreSQL and Redis. Task: configure a production environment, automatic deployment via GitHub Actions, and basic monitoring.

1. Server preparation: SSH key-based access, deploy user, update packages, install Nginx, PHP 8.3 with extensions, Composer, Node.js 20, PostgreSQL 16, Redis.
2. Web server configuration: Nginx configuration for Laravel: single entry point, static cache for a year, gzip for CSS/JS, block access to hidden files.
3. SSL: Let's Encrypt via Certbot with auto-renewal.
4. Deployment via GitHub Actions: Write a workflow: on push to main — copy code, install dependencies, run migrations, optimize, reload PHP-FPM and Nginx.
5. Monitoring: Install Netdata on port 19999 (firewalled via UFW), set up Telegram notifications for overload.

Result: deployment takes 2 minutes (down from 40 minutes), monitoring shows everything in real time, the client sleeps peacefully. By our estimates, deployment automation saves up to 30,000 rubles per month in downtime.

Why deployment automation cuts time by 10 times?

Manual deployment is a chain of commands where mistakes are easy: forgot to run migrations, didn't update cache, messed up permissions. Automation via CI/CD eliminates human error. For example, GitHub Actions runs a sequence: git pull → composer install → php artisan migrate → npm run build → restart services. The whole process takes 3 minutes instead of 30 when done manually. Plus, each step is logged, and in case of an error, you get a notification.