My background is mainly in full-stack web development, especially web frameworks and Node.js. I recently accepted a software engineering position at a defense company where I'll likely work with Docker, Kubernetes, and Helm. I have about one month before I start. Should I focus on learning a broad range of tools or go deeper on the fundamentals? Would my time be better spent pursuing a certification or building and deploying small applications? I've already containerized a frontend and Node API, deployed them to a local Minikube cluster with separate Services and Deployments, used a PersistentVolumeClaim and namespace, and packaged everything in a Helm chart. I can build new images, push them to a registry, update Helm values, and perform upgrades and rollbacks, but I'm concerned that my understanding is still superficial. What should I focus on next?
5 Answers
You probably aren’t expected to arrive as a fully independent DevOps specialist. Since you were hired for a software engineering role, your web and Node experience is likely the main reason they chose you, while the infrastructure skills can grow on the job. Try to learn the tools and patterns they actually use instead of spending the month studying unrelated cloud services or collecting certificates. Build a small project that demonstrates a complete, repeatable deployment and be ready to explain every part of it.
Because this is a defense environment, ask yourself how the same workflow operates without unrestricted internet access. Learn how images and charts are stored in private registries, how dependencies are mirrored, and how configuration and secrets are handled. Practicing in an air-gapped or restricted setup is difficult at home, but understanding the constraints will help. Also review basic Linux, networking, DNS, TLS, and process troubleshooting, since those skills are what you’ll rely on when a deployment fails.
Prioritize hands-on work over a certification right now. Take one small application and run it through the full lifecycle: containerize it, publish the images, deploy it to kind, k3d, or Minikube, expose it through an Ingress, configure it with ConfigMaps and Secrets, and practice upgrades, rollbacks, logging, and troubleshooting. Intentionally break things and repair them. That experience will transfer much better to your first weeks on the job than simply memorizing certification material.
Go deeper on the Kubernetes fundamentals first: kubectl, Pods, Deployments, Services, namespaces, storage, configuration, networking, and basic scheduling. Then make sure you understand what Helm is actually rendering and how releases, values, upgrades, and rollbacks work. GitOps tools such as Argo CD are worth learning at a high level, but trying to master every surrounding tool in one month could spread you too thin.
I agree with this order. A first month is better spent becoming comfortable with core Kubernetes and Helm, then understanding the basic GitOps workflow rather than attempting to become an expert in Argo CD immediately.
A useful next project would be to add observability and failure scenarios to your existing application. Set up health checks, resource requests and limits, rolling updates, persistent storage recovery, logs, and basic metrics. Then test cases such as a bad image, unavailable dependency, failed rollout, full volume, and misconfigured Service. That will push you beyond simply getting the happy path deployed.
That’s mostly what I’ve been doing already, but I still feel like I’m only scratching the surface. I’ll spend more time on troubleshooting and day-two operations.