Production-Grade Kubernetes Platform

A production-ready Kubernetes platform running on Kind, featuring GitOps with ArgoCD, full observability stack, and a demo service.

---

Local Kind Cluster

Prerequisites

• kind – Kubernetes in Docker
• kubectl – Kubernetes CLI
• helmfile – Declarative Helm chart management

Increase Local Resources

Required due to known inotify limits:

sudo sysctl fs.inotify.max_user_watches=524288
sudo sysctl fs.inotify.max_user_instances=512

Cluster Management

Action	Command
Bootstrap cluster	make kind-up
Destroy cluster	make kind-down

---

Services

Demo Service

Property	Value
URL	http://demo.localhost
Purpose	Simple HTTP service demonstrating deployment, health checks, autoscaling, and observability

Endpoints:

Endpoint	Description
/	Main endpoint
/healthz	Liveness probe
/readyz	Readiness probe

Quick Test:

curl -H "Host: demo.localhost" http://localhost/

---

ArgoCD

Property	Value
URL	http://localhost:8080
Username	admin

Note: ArgoCD is intentionally not exposed via ingress locally. Access via port-forward.

Setup Access:

kubectl -n argocd port-forward svc/argocd-server 8080:80

Retrieve Password:

kubectl -n argocd get secret argocd-initial-admin-secret \
  -o jsonpath='{.data.password}' | base64 -d && echo

Manages:

• Platform components
• Demo service
• Observability stack
• CRDs and ordering via sync waves

---

Grafana

Property	Value
URL	http://grafana.localhost
Username	admin
Password	admin

Included Dashboards:

• Demo Service dashboard (SLIs via ingress-nginx)
• Default Kubernetes / node dashboards

---

Prometheus

Property	Value
URL	http://prometheus.localhost

Purpose:

• Inspect metrics
• Validate scrape targets
• Debug PromQL queries

---

Alertmanager

Property	Value
URL	http://alertmanager.localhost

Purpose:

• View active alerts
• Inspect alert routing (local setup)