CONTROL PLANE
k8s-master-01
The coordinator: Kubernetes API, scheduling, cluster state, and encrypted etcd snapshots.
✦ ILKER TAYSI’S DEVOPS PLAYGROUND
Small servers.
Big possibilities.
A little corner of the internet, built at home. A place to sharpen DevOps skills, turn curious ideas into small apps and games, and share them just for the fun of it.
Home-hosted.
Cloud-connected. Always learning.
03nodes, one home cluster
64 GiBcombined node memory
03environments: preview, dev, prod
∞ideas worth experimenting with
01 / THE HOMELAB
A real platform.
A personal playground.
Three Ubuntu nodes run a kubeadm Kubernetes cluster. Development, temporary previews, and public production live side by side, with separate namespaces, guarded access, and a release path that keeps the tested image intact.
WORKER 01
k8s-node-01
A home for containers, application replicas, monitoring, and persistent workloads.
WORKER 02
k8s-node-02
A second worker spreads production replicas and leaves room to keep experimenting.
A shared home cluster keeps things approachable. Namespace boundaries, quotas, restricted pods, and backups make the experiments deliberate.
02 / THE TOOLBOX
A thoughtful stack,
from bare metal to browser.
Tools with a purpose: reproducible infrastructure, predictable releases, and enough visibility to understand what’s happening.
Foundation & infrastructure
The hosts and the infrastructure that make everything repeatable.
Orchestration & networking
Containers, cluster networking, and an entry point for every app.
Code & continuous delivery
Build and test once. Let Git describe what should be running.
Edge & access
A public front door for projects, with protected doors for development.
Observability & upkeep
Metrics inside the lab. Uptime checks outside the home network.
Data & recovery
PostgreSQL, persistent storage, and tested offsite recovery paths.
03 / UNDER THE HOOD
One home. A few well-connected layers.
Follow a request, a release, or a backup through the platform.
Cloudflare is the front door. Public apps reach Envoy through an outbound tunnel; development, previews, and admin tools require Access.
Observability runs alongside the workloads: Prometheus collects metrics, Grafana visualizes them, and UptimeRobot watches public apps from outside the lab.
04 / FROM IDEA TO INTERNET
A small change.
A complete delivery cycle.
Choose a step to see how a project moves from an idea to a public release. Production gets the image that development already tested.
THE STARTING POINT
An idea becomes an application.
A scaffold creates a static site, API, or Next.js app with a Dockerfile, a health endpoint, and a place in the application registry. Prepare the namespace, then make it your own.
Ubuntu / WSL
git switch -c pv/myapp/first-release origin/develop
python3 scripts/scaffold.py myapp static
python3 ../infra/scripts/prepare_namespace.py dev myappOne branch. One app. Plenty of room to experiment.
05 / OUT IN THE WORLD
Little projects,
made to be shared.
Apps, experiments, and eventually a few games. This shelf discovers public production deployments automatically—there’s always room for the next idea.