11 Independent DevOps Learning Sites That Are Not Selling You a Certification
Search for how to learn DevOps and you get two kinds of result. The first is vendor content. It is often excellent, and it will teach you the vendor's product. You will come away knowing how their managed service handles a thing, and not much about what the thing is. The second is certification training. Also often good, and optimised for passing an exam. That is a legitimate goal, but the exam is a proxy, and the proxy has drifted. There is a third category that is harder to find because nobody buys ads for it: resources built by individuals and small teams because they wanted the thing to exist. They rarely have a marketing budget. Some are ugly. They are the ones I keep going back to. Here are eleven, grouped by what they are actually for. I checked each one was still alive before writing this, which is not a given for lists like these. 1. DevOps Daily What it is: DevOps Daily is an independent site with written guides, interactive simulators and quizzes, plus tool comparisons for the parts of the stack where the choice actually matters. What it is good for: the gap between reading about a tool and having used one. The simulators let you work through a scenario in the browser, and the comparisons are written for the moment you have two options open in tabs and need to know what the actual trade-off is. Where it stops: it is newer and thinner than most of what follows. Coverage is uneven, some areas are a single guide deep, and there is no community around it yet. Judge it against the rest of this page rather than because it is at the top. For getting your bearings 2. roadmap.sh What it is: community-maintained learning roadmaps as interactive diagrams. The DevOps roadmap lays out the whole field as a dependency graph, so you can see what depends on what. What it is good for: the moment you realise DevOps is not a subject but about nine subjects, and you have no idea which to start with. The roadmap answers that specific question better than anything else I have used. Where it stops: it tells you what to learn, not how. It is a map, not a course, and treating it as a checklist to grind through is how people end up with shallow familiarity in twelve areas. 3. The Twelve-Factor App What it is: 12factor.net, a document from 2011 about how to build applications that can be operated. What it is good for: it is the shortest path to understanding why a lot of modern infrastructure looks the way it does. Config in the environment, stateless processes, logs as event streams. Almost every container practice traces back to something in here. Where it stops: it is fifteen years old and some of it shows. Read it as foundations rather than current practice. For actually doing things 4. SadServers What it is: SadServers gives you a broken Linux box in the browser and a problem statement. Disk full, service will not start, mystery process eating CPU. You fix it. What it is good for: the single biggest gap between people who have done this job and people who have studied it is troubleshooting under uncertainty. This is the only resource I know that drills that directly. It is also the best interview preparation available for infrastructure roles, because the scenarios are shaped exactly like the interview. Where it stops: scenarios are individual boxes. It will not teach you distributed systems, and it is deliberately not a course. 5. Killercoda What it is: Killercoda runs interactive scenarios in real environments in the browser. It picked up much of the ground Katacoda left when O'Reilly retired it. What it is good for: trying something without installing anything. Real Kubernetes clusters that exist for twenty minutes and then vanish, which removes the single biggest barrier to practising: not wanting to break your laptop. Where it stops: scenario quality varies because much of it is community-contributed. The official tracks are consistently good; the long tail is a mixed bag. 6. OverTheWire: Bandit What it is: Bandit is a wargame. You SSH into a box and each level's password is hidden somewhere you need a shell skill to reach. What it is good for: genuine command-line fluency, learned the only way it sticks, which is by needing it. Thirty-odd levels take you from ls to reasonably comfortable with find, grep, pipes, permissions and compression. Where it stops: it is framed as security, so people in infrastructure skip it. That is a mistake. The first twenty levels are pure Linux fundamentals. 7. Container Training What it is: container.training, Jérôme Petazzoni's complete Docker and Kubernetes courseware. Full slide decks, exercises and instructor notes, free, from someone who was at Docker early. What it is good for: depth. This is material from multi-day paid workshops, published in full. If you want to understand container internals rather than commands, start here. Where it stops: it is courseware, so it reads like courseware. You are teaching yourself from an instructor's notes, which takes more effort than a tutorial. For understanding rather than doing 8. Wizard Zines What it is: Julia Evans' zines on networking, debugging, containers, bash, DNS. Illustrated, short, and unreasonably clear. The free comics alone are worth the visit. What it is good for: the specific feeling of using something for years without understanding it. Her explanations of DNS and of how containers actually work are better than the official documentation for either. Where it stops: the good stuff is paid, though fairly priced and the free material is substantial. 9. Kubernetes The Hard Way What it is: Kelsey Hightower's walkthrough of standing up a cluster from components, no install scripts. What it is good for: it is the fastest way to stop treating Kubernetes as magic. You generate the certificates, start the components, wire them together, and the architecture stops being abstract. Where it stops: the title is honest. It takes a day and you will never build a cluster this way in production. That is the point, and people who skim it get nothing. 10. Brendan Gregg's site What it is: brendangregg.com, decades of performance analysis writing from someone who spent years at Netflix doing exactly this. What it is good for: the day something is slow and nobody knows why. His USE method gives you a procedure to follow instead of guessing, which is the difference between debugging and poking. Where it stops: it is dense and assumes you already know what a system call is. Not a starting point. 11. explainshell What it is: explainshell takes a shell command and annotates each flag with the relevant chunk of the man page. What it is good for: the twenty seconds after you copy a command off Stack Overflow and want to know what -xzvf is doing before you run it as root. That habit alone makes it worth bookmarking. Where it stops: it is a lookup tool, not a learning path. But it is the one on this list I use most often. What I left out, and why Vendor platforms. Play with Docker, cloud provider training, most managed-service tutorials. Frequently high quality, and all of them have a product at the end of the path. Worth using with that in mind, but not what this list is about. Certification courses. If you need the certificate for a job filter, buy the course that gets you the certificate. That is a rational purchase and a different question from learning. Katacoda. It appears on every list like this and O'Reilly retired it in 2022. The domain still responds, which is presumably why the lists never got updated. Anything I could not verify. A couple of well-regarded independent sites did not resolve when I checked. Rather than list them from memory and send you to a dead domain, I left them out. The pattern worth noticing Nearly all of these were built by one person, or a handful, scratching their own itch. None of them is trying to move you along a funnel. That is also their weakness. They are unevenly maintained, sometimes ugly, and no one is going to hold your hand. But the understanding you get from generating Kubernetes certificates by hand, or from fixing a genuinely broken server with no hints, is a different quality of understanding from watching someone else do it. What have I missed? I am specifically after independent ones, not the big platforms.
This is a summary aggregated from Dev.to. Read the complete article on the original site:
Read full article at Dev.to