Kubernetes Infrastructure Project

Automated Kubernetes infrastructure on Debian Bookworm, managed with Ansible. Deploys a single-node K8s cluster with monitoring (Prometheus, Grafana, Alertmanager), Nginx web/proxy servers, a 5-node etcd cluster, and a health-checker daemon.

Prerequisites

• Debian Bookworm (amd64)
• ansible installed (sudo apt-get install -y ansible)
• Root or sudo access

Quick Start

sudo apt-get install -y ansible
sudo ansible-playbook playbooks/site.yml

Services

Service	NodePort	Notes
Prometheus	30090	Metrics & alerting
Grafana	30030	Dashboards (admin/admin)
Alertmanager	30093	Alert routing
Nginx Web	30081	Static webserver
Nginx Proxy HTTP	30080	Redirects to HTTPS
Nginx Proxy HTTPS	30443	SSL/TLS reverse proxy + cache
Etcd	30379	5-node etcd cluster

Verification

# Check all pods
kubectl get pods -A

# Nginx webserver
curl http://localhost:30081/

# Nginx proxy (HTTPS with self-signed cert)
curl -k https://localhost:30443/

# Prometheus
curl http://localhost:30090/-/ready

# Alertmanager
curl http://localhost:30093/-/healthy

# Etcd health
curl http://localhost:30379/health

Fault Tolerance Test (Etcd)

The etcd cluster runs 5 nodes and tolerates 2 simultaneous failures.

# Scale down to 3 nodes (simulate 2 failures)
kubectl scale statefulset/etcd -n database --replicas=3

# Verify etcd still works with quorum
kubectl exec -n database etcd-0 -- etcdctl \
  --endpoints=http://etcd-0.etcd-headless.database.svc:2379 \
  put /fault "ok"

# Scale back to 5
kubectl scale statefulset/etcd -n database --replicas=5

Requirements Checklist

• Docker/Kubernetes containers on Debian Bookworm
• Prometheus + node-exporter + nginx-exporter metrics
• BONUS Grafana dashboards
• BONUS2 Alertmanager with alert rules
• Nginx webserver (NodePort 30081)
• Nginx reverse proxy + cache (NodePort 30080/30443)
• BONUS Keepalive upstream connections
• BONUS2 SSL/TLS (TLSv1.2+, strong ciphers, HSTS)
• Etcd 5-node cluster (tolerates 2 failures)
• Daemon with daemontools (svscanboot + svstat)
• BONUS multilog log rotation
• Load metrics + nginx stub_status collection
• Human-readable output format (key=value pairs)
• Ansible automation (single ansible-playbook command)
• Single-node Kubernetes cluster (kubeadm)