Sunday, 06 September 2026
Advertisement Advertise Your advert could be here Reach thousands of learners and ICT professionals across Rwanda. Contact us
Advertisement Opportunity Jobs, scholarships & hackathons Fresh openings from Rwandan job boards are pulled in every hour. See openings

What DevOps actually is

DevOps from zero · lesson 1 of 12

In this lesson: Explain in one sentence what DevOps is for, without using the word "culture".

DevOps is the practice of making the path from "a developer changed something" to "users are safely using it" short, repeatable and boring. That is the whole idea. Everything else — the tools, the diagrams, the job titles — exists to serve that sentence.

The problem it was invented to solve

For a long time, software companies were split in two.

Developers were paid to change things. A new feature, a fix, a redesign. Their success was measured in how much shipped.

Operations were paid to keep things running. Their success was measured in uptime. And the single most reliable way to keep a system up is to never change it.

So the two halves of the same company were paid to want opposite things. The result was predictable and it had a name: throwing it over the wall. Developers finished a feature, handed over a folder and some instructions, and moved on. Operations received software they had never seen, deployed it at night, and were woken at 3am when it broke. Nobody could fix it: the people who understood the code were asleep, and the people who were awake had no idea what it did.

The insight: the problem was never that operations people were too cautious or that developers were too careless. It was that the organisation had given them conflicting goals and then acted surprised when they conflicted.

What changed

DevOps removes the wall. The same team writes the software and is responsible for it running. If you write it, you deploy it; if it wakes someone at 3am, it wakes you. That sounds like a punishment. In practice it is the opposite: when the person who can fix the design is the person being woken up, the design improves very quickly.

But you cannot ask one team to do both jobs unless deployment stops being an event. If shipping a change takes four hours of careful manual work, nobody will do it often, and every release becomes big and risky. So DevOps in practice means automating the path to production until running a release is as unremarkable as saving a file.

The four things you are actually building

GoalWhat it means in practice
RepeatableThe same steps run the same way every time, whether it is Monday morning or Friday night.
FastA small change reaches users in minutes, so problems are found while the author still remembers the code.
ObservableWhen something breaks you can find out what, without guessing.
ReversibleA bad release can be undone quickly. This is what makes speed safe.

Small and often beats big and rare

This is the least obvious idea in the whole field, so it is worth stating plainly. Teams that deploy many times a day have fewer outages than teams that deploy once a quarter, not more.

The reason is arithmetic. If you ship one change and it breaks, you know exactly what broke it. If you ship two hundred changes at once and something breaks, you have two hundred suspects and no way to tell them apart. The quarterly release is not safer — it is just a bigger accident waiting to happen, with a longer investigation afterwards.

What this course will and will not do

It will teach you the ground floor: the shell, Git, how a request reaches your application, how to run and look after a real server, and how to read what a machine is telling you. Those skills are not glamorous and they do not change every year, which is exactly why they are worth having.

It will not teach you Kubernetes. That comes later, and starting there is like learning to fly a plane before you can drive. Almost everything that goes wrong at large scale is a familiar small problem repeated a thousand times.

Try it yourself

Think of any process you already run by hand — closing the books at month end, setting up a new employee's accounts, publishing a newsletter. Write down every step. Now mark each one: could a machine do it, or does it genuinely need a human decision? The steps you marked "a machine could do this" are exactly the kind of work DevOps removes. Keep the list; you will recognise the same shape when we get to pipelines.

Create a free account to save progress

All lessons in this track

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
    Your first server: SSH and keys ~35 min account needed
  8. 8
    Deploying by hand, and why it hurts ~35 min account needed
  9. 9
  10. 10
  11. 11
  12. 12
    CI/CD in plain words ~25 min account needed
Advertisement Yanjye Learn a new digital skill this week ICT, programming and professional courses with graded weekly assignments. Start free