A good DevOps course in Nepal makes you break and fix real servers every week, and it tells you who pays the cloud bill. Ask for the week-by-week syllabus before you pay, and check six things in it.

If a course cannot show you that syllabus, you already have your answer. The questions below work for any course, ours included. They come from what the job asks for: set up a server, secure it, send new code to it, watch it and fix it when it breaks.

What should the first weeks of a DevOps course cover?

Linux, networking and Git. Not Kubernetes.

Most DevOps problems end in a terminal. A pod keeps crashing, a deploy fails, a site will not load. You find the cause by reading logs with journalctl, grep and tail, and by checking DNS, ports and the firewall in order. A learner who skipped these weeks can copy a Kubernetes file but cannot say why it fails.

So open the syllabus and look at the first month. If Docker and Kubernetes appear in week 1 or 2, ask what happens to someone who has never used a terminal. A clear answer names the Linux topics and the weeks they take.

Who pays the AWS bill, and how do you avoid a surprise charge?

Most people forget this question. It matters more when you pay with a dollar card from a Nepali bank.

AWS now gives new accounts a Free plan. You get USD 100 in credits when you sign up, and you can earn up to USD 100 more. The account closes on its own after 6 months or when the credits run out, whichever comes first. AWS says you are not charged unless you move to a Paid plan.

That is enough for a course, if the course is careful. Some services cost money every hour they exist, even when idle. An Amazon EKS cluster, for example, costs USD 0.10 per hour, which is about USD 73 a month before any servers. Ask the course which plan the labs use, which costly services they avoid, and whether you delete everything at the end of each lab.

Do you fix broken things, or only build new ones?

Building is the easy half. Employers hire juniors who can set up a server and also find out why it stopped.

Ask how often you debug something the mentor broke on purpose. Ask what you use to find the cause. A good answer sounds like "every week, from the logs, one step at a time". A weak answer is a list of tools with nothing about failure.

How does the course handle passwords and keys?

A leaked cloud key is the fastest way for a beginner to get a large bill. So ask how the CI/CD pipeline logs in to the cloud. CI/CD means the automatic build, test and deploy that runs on every change.

The safer way is OpenID Connect, or OIDC. GitHub's own docs explain it. The workflow gets a short-lived token from the cloud provider, valid for a single job, so no long-lived key sits in GitHub secrets. A course that still has you paste access keys into a repository is teaching an old habit.

Where do you run Kubernetes?

You do not need a paid cloud cluster to learn Kubernetes. k3d runs k3s, a small Kubernetes, inside Docker on your own machine. Ask whether the course does this, or puts you on cloud time that someone pays for.

Ask the laptop question too. Docker plus a local cluster needs memory. Find out how much RAM the course expects, and sort it out before the first class, not during it.

What job does the course say you will be ready for?

Be careful with any course that promises a "DevOps Engineer" title at the end. Our own syllabus says it plainly: most DevOps jobs ask for 2 or more years of experience. A real first step is an internship or a junior cloud support job, working under a senior on tickets, monitoring and pipelines.

Ask what the course writes about this. Then ask what you will be able to show: projects on GitHub that a person reviewed, a CV built around them, and practice in a troubleshooting interview.

Cloud and DevOps jobs in Nepal are posted on company career pages, merojob, KumariJob, JobAxle and LinkedIn. Read ten posts there before you pay. Then check the syllabus against the tools those posts name.

How Saarathi Academy answers these questions

Saarathi Academy in Old Baneshwor, Kathmandu runs a 10-week Cloud Computing and DevOps course, 2 hours a day, in person and online. Up to 10 students sit in the classroom per batch, and mentors still do the job they teach. Here are our answers, from the public syllabus:

  • Weeks 1 to 4 are Linux, Bash, Git, networking and a safe AWS account. Docker starts in week 5.
  • Labs stay on the AWS Free plan, with budget alarms at USD 1, 5 and 20, and terraform destroy is graded at the end of every lab.
  • EKS is shown by the mentor only. You run Kubernetes on your laptop with k3d, then on one small server you tear down after.
  • In weeks 1 to 9, the mentor breaks something on purpose every week, and you find the cause from the logs.
  • Pipelines log in to AWS with OIDC, never a stored key.
  • The goal is an internship or a junior cloud support job, with no salary promise. You need a laptop with at least 8GB of RAM.

Try this tonight

If you have Docker installed, install k3d and run k3d cluster create mycluster, then kubectl get nodes. You now have a Kubernetes cluster that costs nothing.

If you already have an AWS account, open Budgets and create a Zero spend budget. It emails you once your spending goes past the free limits.

Our full week-by-week plan is on the course page and in the syllabus.