I've worked in Linux application support for about 800 days, mostly monitoring disk usage, checking application logs, and supporting services running on Rocky Linux and Red Hat systems. I can deploy simple applications such as Vaultwarden quickly with Docker Compose, but I don't feel confident designing or securing a production-grade environment.
I understand that production deployment involves more than copying a Compose file, but I'm unsure what concepts I should learn next. I've built toy servers and followed courses on tools like Kubernetes, but I still feel like I'm relying on trial and error without understanding the bigger picture. I'd like to develop the skills needed to deploy, operate, secure, and troubleshoot real production services instead of remaining stuck in application support.
3 Answers
Start with the fundamentals rather than collecting tools. Build a strong base in Linux administration, networking, storage, permissions, authentication, backups, monitoring, logging, and incident response. Server hardening is only one part of production readiness.
The best way to learn is to run a small service and document everything: how it is deployed, what happens when it crashes, how backups are restored, how certificates are renewed, how access is controlled, and how you would investigate an outage. Add automation and infrastructure-as-code once you understand the manual process. Kubernetes, Helm, Argo CD, and Terraform are useful later, but they won’t replace the underlying concepts.
Your current title may be limiting your growth. If your work is mostly checking logs and supporting applications, you may have useful operational experience but not much ownership of infrastructure or releases. Look for opportunities to automate repetitive tasks, improve monitoring, participate in incident reviews, write deployment runbooks, and own a service through its full lifecycle.
There are established resources on this subject, including the Google SRE books. They cover reliability, service-level objectives, monitoring, incident management, and designing systems that can be operated safely. You don’t need to master every advanced platform immediately; aim to understand reliability and operations first, then choose tools that solve specific problems.
That makes sense. I’ve been treating DevOps as a list of technologies to study, which has left me with a lot of tutorials but little confidence. Focusing on ownership, reliability, and repeatable operations seems like a better direction.
Deployment is not just getting a container to start. In production you also need repeatability, secure configuration, least-privilege access, patching, secrets management, backups, observability, rollback procedures, capacity planning, and a way to recover from failure.
Try creating a complete lab project instead of isolated toy servers. For example, deploy an application with a reverse proxy, TLS, a database, restricted firewall rules, centralized logs, metrics and alerts, automated backups, and a documented restore test. Then deliberately break parts of it and practice recovery. That will teach you much more than repeatedly installing individual technologies.
That’s the part I’ve been missing. I can follow tutorials for Kubernetes or Terraform, but I don’t always understand why each component is needed. I should probably focus on designing and operating a smaller system end to end before adding more tools.