1. I thought the next step after Excel was a website. Why is there suddenly a logistics network?
Someone who has mostly used spreadsheets enters web development and asks a perfectly reasonable question:
Wasn’t I just trying to publish a site?
HTML is understandable.
Then the following arrive at once:
- Git
- GitHub
- repository
- branch
- commit
- build
- deploy
- Worker
- D1
- KV
- DNS
- CDN
- CI/CD
You leave the starting village and instead of a slime, “Distributed Systems 101” is waiting.
The confusion has a structural cause. Spreadsheet work often stays inside one application. A web service separates writing, storing, transforming, executing, delivering, and recording data into different systems.
2. The real problem is not difficulty. It is skipped layers
There is no single universal curriculum, but conceptually web development often accumulates layers:
- write files
- track changes with Git
- store and collaborate through GitHub
- build the project
- deploy it
- run APIs
- connect a database
- automate operations
- handle retries and failures
Modern templates and AI can push a beginner from step 1 to step 9 very quickly.
That is convenient.
Your head can barely keep up.
One day you learn “GitHub stores code.” The next day you hear “Workers Builds root,” “D1 binding,” and “KV is eventually consistent.”
That is like learning to use a refrigerator and being immediately assigned to food-factory logistics.
3. What is GitHub? Mostly a blueprint warehouse plus history
GitHub’s central job is to manage files in Git repositories.
A website repository might contain articles, HTML, CSS, JavaScript, configuration, tests, and scripts.
Commits record what changed and when.
A useful rough model is:
GitHub = blueprint warehouse + change history
GitHub can also host static sites through GitHub Pages and automate builds, tests, and deployments with GitHub Actions.
That overlap is what makes beginners suspicious.
The warehouse also contains a small factory.
But GitHub’s center of gravity is still code, collaboration, and change management.
4. What is Cloudflare? The factory suddenly owns the store, warehouse, trucks, and security
Cloudflare can do many jobs:
- serve websites
- run server-side code with Workers
- build and deploy projects
- store SQL data with D1
- store key-value data with KV
- distribute content through a global network
- manage DNS
- provide security controls
- run durable multi-step processes with Workflows
Calling it merely “hosting” misses the scale of the toolbox.
A better image is:
factory + store + warehouse + delivery network + security room
And yes, the database is on the same industrial estate.
GitHub and Cloudflare overlap in some areas, but their main roles differ.
5. What is a build? Turning the blueprint into a product
“Build” sounds abstract until you translate it as assembly.
Files in a Git repository are often not the finished website.
With a framework such as Astro, content, components, configuration, and code are transformed into distributable HTML, CSS, and JavaScript.
That transformation is the build.
Cloudflare Workers Builds can react to a push in a connected Git repository, run a build command, and then run a deploy command.
So:
GitHub = model kit box
Build = assembly
Deploy = put the finished item on the shelf
Build logs are long because they are factory logs.
You wanted the website.
The system handed you the manufacturing report.
6. D1 and KV: both store data, so why are there two?
D1 is Cloudflare’s SQL database. It stores data in tables and uses SQLite-style SQL: a way to ask for rows that match a condition.
Think of it as an official ledger with rows, columns, queries, and relationships.
KV stores a name and its value together. For example, the name “site mode” can have the value “normal.”
Examples:
- site_mode → normal
- feature_x → true
- user_123_theme → dark
It behaves more like a giant dictionary.
KV is optimized for fast reads and global caching, but it is eventually consistent. Cloudflare documents that another location may temporarily see an older value, sometimes for 60 seconds or more.
So a useful model is:
D1 = official ledger
KV = fast dictionary or sticky-note board
You may reasonably ask why everything cannot be one database.
The answer is that distributed systems enjoy creating specialized drawers.
7. Which one is better at running by itself?
It depends on the job.
GitHub Actions can automatically run jobs such as build, test, and deploy when repository events occur.
Cloudflare can run APIs continuously with Workers. Workflows remembers which steps are finished, tries failed steps again, waits when needed, and picks up where it left off.
As a metaphor:
GitHub = engineer good at design, change, and repair
Cloudflare = automated factory good at executing a defined process
If the question is “What code changed and how should we fix it?”, GitHub is the natural center.
If the question is “Keep this process moving and retry failures until the next step,” Cloudflare is the natural center.
Different jobs, not a simple winner.
8. When does everyone learn this? They do not
GitHub is common in computer science courses, research labs, and software development jobs.
Cloud platforms, databases, CI/CD, and products such as Cloudflare are often learned only when someone needs to deploy and operate real services.
A non-engineer can spend an entire career without touching Cloudflare Workers.
Many jobs in economics, accounting, sales, HR, and operations may use spreadsheets heavily while never requiring infrastructure tools.
So the correct interpretation is not:
“Everyone knew this except me.”
It is:
“I opened another profession’s toolbox.”
Someone can study marginal utility and later get hit by “eventual consistency.”
That is not a normal curriculum transition.
9. Conclusion: you do not need to memorize everything. Learn who owns which job
The first useful map is small:
| Name | One-line meaning |
|---|---|
| GitHub | blueprints and history |
| Build | turn source into a finished artifact |
| Deploy | place the artifact in production |
| Cloudflare Workers | server-side execution |
| D1 | SQL ledger |
| KV | fast key-value dictionary |
| GitHub Actions | GitHub-side automation |
| Cloudflare Workflows | long jobs that remember where they left off |
Once this map exists, an error becomes easier to classify:
GitHub problem?
Build problem?
Worker problem?
Database problem?
That classification alone removes a lot of noise.
The overload did not happen because you were too slow.
It happened because you moved from one-room spreadsheet software into the facility-management room of an entire shopping mall.
You wanted to change a light bulb.
Now someone is explaining the fire shutters.
References (7)
- GitHub Docs, What is GitHub Pages? https://docs.github.com/en/pages/getting-started-with-github-pages/what-is-github-pages
- GitHub Docs, Understanding GitHub Actions https://docs.github.com/en/actions/get-started/understand-github-actions
- Cloudflare Docs, Workers Builds https://developers.cloudflare.com/workers/ci-cd/builds/
- Cloudflare Docs, Workers Builds Configuration https://developers.cloudflare.com/workers/ci-cd/builds/configuration/
- Cloudflare Docs, D1 https://developers.cloudflare.com/d1/
- Cloudflare Docs, How KV works https://developers.cloudflare.com/kv/concepts/how-kv-works/
- Cloudflare Docs, Workflows https://developers.cloudflare.com/workflows/
