field diary · 12 chapters · free, no signup
Learn DevOps by walking one pipeline, from ls to a canary release.
Every topic gets three passes: a plain-English explanation, a real-life analogy you already understand, and a lab you can actually run on your own laptop. Then you play with it until it sticks.
no accounts · no cloud bill · runs on Windows, macOS or Linux
how it works
Three passes over every single topic.
Most DevOps material either drowns you in vocabulary or gives you commands to copy with no idea what they do. Each topic here is written the way a patient senior would explain it at a whiteboard.
pass one
Plain language
What the thing is, what problem it was invented for, and where it sits in the pipeline — in short sentences, with the jargon defined the first time it appears.
pass two
A real-life analogy
Every abstract idea is mapped to something you have already lived through: a tiffin box, a kitchen during dinner rush, a hostel water tank, an airport security queue.
pass three
Hands on the keys
A short lab with the exact commands, the output you should expect, and one thing to deliberately break — because you only really learn a system by watching it fail.
sample · chapter 07, topic 3
What a lesson looks like
This is a real excerpt, not a mockup. Docker images are one of the ideas people nod along to for months without ever picturing correctly, so it gets an analogy, a diagram in words, and a command whose output you can compare against your own screen.
An image is a recipe printed on a card. A container is the dish you cooked from it tonight. You can cook the same card a hundred times in a hundred kitchens and get the same dish — and throwing away tonight's dish never damages the card.
Each layer of an image is one more line written on that card. Change the last line and you only re-cook the last step; change the first line and everything after it has to be made again. That is the whole secret to fast builds.
docker build -t notes-api:v1 .
=> [1/5] FROM node:20-alpine CACHED
=> [2/5] COPY package*.json ./ CACHED
=> [3/5] RUN npm ci CACHED ← 42s saved
=> [4/5] COPY . . 0.2s
=> [5/5] RUN npm run build 3.1s
=> exporting layers 0.4s
Successfully tagged notes-api:v1
docker run -d -p 3000:3000 notes-api:v1
c3f9a1b2d4e5
curl -s localhost:3000/health
{"status":"ok","uptime":1.42}
Move one line and watch the cache die
Put COPY . . above RUN npm ci, rebuild, and time it. Every code change now re-installs every dependency. That one-line mistake is the most common reason a team's pipeline takes nine minutes instead of one.
the syllabus · read in order, or jump to what you need
Twelve log entries, beginning to on-call.
the arcade · seven games, no install
Reading teaches the words. Playing teaches the reflexes.
Each game drills one thing that beginners get wrong in interviews and in production. They run in this browser, keep score, and tell you why an answer was wrong.
Pipeline Builder
Drag eight stages into the order a real deployment runs. Put the security scan after the deploy and the game will tell you exactly who you just paged.
Terminal Trainer
You get the goal — "show the last 50 lines of the nginx error log and follow it" — you type the command. 30 tasks across Linux, Git, Docker and kubectl.
Container or VM?
Sixteen statements, two bins. The fastest way to stop mixing up the two things every DevOps interview opens with.
YAML Doctor
A manifest is broken and the cluster is refusing it. Click the line that is wrong. Five patients, increasingly sneaky.
Incident Room
3:14 a.m., checkout is failing, you are on call. Every choice moves the clock. Try to close it in under 20 minutes without making it worse.
Tool Match
Memory game: pair each tool with the job it was built for. Ends the habit of saying "we'll use Kubernetes" for a static site.
Interview Blitz
Forty-one rapid questions with a timer and a streak counter — the ones that actually come up in a first-round DevOps interview.
a plan that fits a job
Four weeks, an hour a day.
If you have more time, do the labs twice — once following along, once from memory. The second run is where the learning happens.
week one
Ground floor
Chapters 01–04. Get comfortable in a terminal and stop being afraid of Git. Finish by pushing a branch, opening a pull request and merging it yourself.
week two
Automate the boring
Chapters 05–06. A green pipeline on every push, and your own app running in a container you built. This is the week it starts feeling real.
week three
Run it like production
Chapters 07–09. A local cluster, infrastructure defined in code, and a canary release you can roll back on purpose.
week four
Keep it alive
Chapters 10–12. Dashboards, one alert worth waking up for, a scan in the pipeline, and a portfolio project written up properly.