Situation
The team’s production services run on Linode VPS hosts. Someone has to create them, point domains at them, keep certificates valid, control who logs in, patch and back them up, and look first when a service stops answering. That’s my job next to the pipelines.
What I handle
| Layer | Work |
|---|---|
| Servers | New VPS from scratch (OS, users, PHP-FPM, Node, Docker). Resizing or splitting hosts based on the dashboards, and sizing plans to real usage to keep cost down. |
| Edge | Domains and DNS records, SSL certificates and renewals, Nginx, Apache and Caddy virtual hosts. |
| Access | SSH keys and permissions on servers and platform accounts, granted on joining and removed on leaving. Firewall rules, fail2ban, SSH config and OS patches. |
| Secrets | Per-environment .env files and GitHub Actions secrets, kept out of repositories. |
| Data | PostgreSQL and MySQL users and grants. Database dumps, server snapshots and restores. |
| Processes | PHP-FPM pools, cron and the Laravel scheduler, queue workers and Horizon under Supervisor, log rotation. |
| Docs | The deployment SOPs and workflow docs developers release with. |
When a service is down
I work from the outside in:
- Reachable at all? DNS, certificate, port, web server.
- Process up? PHP-FPM, the container, Supervisor.
- What do the logs say?
- Config, secrets and cached settings correct?
- Can it reach the database and Redis?
For incident handling, see the monitoring case.
Limits
This is VPS operations, not a high-availability or multi-region platform. AWS, GCP and Kubernetes are on my learning path, not in my work experience.