WordPress on Kubernetes with Monitoring

A production-ready WordPress deployment on Kubernetes with custom Docker images, Prometheus monitoring, and Grafana visualization. Features OpenResty (Nginx with Lua) as a reverse proxy, persistent storage for scaling, and comprehensive metrics collection.

Features

• Custom Docker Images: WordPress, MySQL, and OpenResty built from Dockerfiles
• OpenResty with Lua: Nginx compiled with Lua support for advanced routing
• ReadWriteMany Volumes: Persistent storage supporting horizontal scaling
• Prometheus Monitoring: Real-time metrics collection and alerting
• Grafana Dashboards: Visualization of nginx requests, error rates, and pod CPU
• Helm Charts: Easy deployment and management

Prerequisites

• Kubernetes cluster (minikube, kind, or cloud provider)
• kubectl configured to access your cluster
• Helm 3.x installed

Architecture

Internet → Nginx (OpenResty + Lua) → WordPress (PHP-FPM) → MySQL
              ↓
         Nginx Exporter → Prometheus → Grafana
              ↓
         Pod Metrics

Components:

• OpenResty (Nginx): Reverse proxy with Lua scripting and Prometheus exporter
• WordPress: PHP-FPM application server with persistent storage
• MySQL: Database with persistent volume
• Prometheus: Metrics collection and alerting engine
• Grafana: Metrics visualization with pre-configured dashboards

Quick Start

1. Deploy WordPress Stack

# From the project root
helm install my-release ./charts/wordpress

This creates:

• WordPress deployment with custom Docker image
• MySQL database with PersistentVolumeClaim (ReadWriteMany)
• OpenResty (Nginx) with Lua support and metrics exporter
• Persistent volumes for WordPress content and database

Verify deployment:

kubectl get pods
kubectl get pvc
kubectl get svc

2. Deploy Monitoring Stack

# Add Prometheus community Helm repository
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update

# Install Prometheus and Grafana
helm install monitoring prometheus-community/kube-prometheus-stack

Wait for all pods to be ready:

kubectl get pods -w

3. Access Services

WordPress Site

# Get the URL (for minikube)
minikube service nginx --url

# Or for cloud deployments
kubectl get svc nginx

Visit the URL in your browser to complete WordPress installation.

Grafana Dashboard

# Port forward Grafana
kubectl port-forward svc/monitoring-grafana 3000:80

• URL: http://localhost:3000
• Username: admin
• Password: Get with this command:

kubectl get secret monitoring-grafana -o jsonpath="{.data.admin-password}" | base64 --decode

Prometheus UI

kubectl port-forward svc/monitoring-kube-prometheus-prometheus 9090:9090

• URL: http://localhost:9090

Verify Data Sources

Grafana comes pre-configured with Prometheus and Alertmanager data sources. To verify:

• Navigate to Configuration → Data Sources in Grafana
• Both Prometheus and Alertmanager should show as connected

Monitoring & Metrics

This project implements comprehensive monitoring as per production requirements:

Monitored Metrics

1. Pod CPU Utilization

• Tracks CPU usage across all pods
• Alert threshold: >80% CPU usage
• Visualized in Grafana dashboards

2. Nginx Request Metrics

• Total Request Count: All HTTP requests through nginx
• Total 5xx Requests: Server error tracking
• Error Rate: Percentage of failed requests
• Alert threshold: >5% error rate

3. Additional Metrics

• Pod memory usage
• Nginx connection count
• MySQL connection status
• Container restart count

Custom Dashboards

Navigate to Grafana → Dashboards to view:

• Total Nginx Requests: HTTP request volume over time
• 5xx Error Rate: Application error monitoring
• Pod CPU Usage: Resource utilization tracking

Alerts

Prometheus alerting rules are configured for:

• HighPodCPU: Triggers when pod CPU usage exceeds 80%
• High5xxErrorRate: Triggers when 5xx error rate exceeds 5%

Check alerts in Prometheus UI → Alerts tab.

For complete metrics documentation including WordPress, Apache, and Nginx metrics, see METRICS.md.

Troubleshooting

Check pod status

kubectl get pods
kubectl describe pod <pod-name>
kubectl logs <pod-name>

Verify services

kubectl get svc
kubectl get servicemonitor

Check metrics endpoint

kubectl port-forward svc/<release-name>-nginx-metrics 9113:9113
curl http://localhost:9113/metrics

MySQL connection issues

# Get MySQL password
kubectl get secret <release-name>-mysql -o jsonpath="{.data.mysql-root-password}" | base64 --decode

# Connect to MySQL pod
kubectl exec -it <mysql-pod-name> -- mysql -u root -p

Cleanup

# Delete WordPress stack
helm delete my-release

# Delete monitoring stack  
helm delete monitoring

# Delete persistent volumes (optional)
kubectl delete pvc --all

Custom Docker Images

This project uses custom-built Docker images located in charts/wordpress/docker/:

OpenResty (Nginx)

• Built from source with Lua support
• Compiled with configure options:

  ./configure --prefix=/opt/openresty \
              --with-pcre-jit \
              --with-ipv6 \
              --without-http_redis2_module \
              --with-http_iconv_module \
              --with-http_postgres_module \
              --with-http_stub_status_module \
              -j8

• Includes nginx-prometheus-exporter for metrics
• Proxy pass configuration from nginx to WordPress

WordPress

• Custom PHP-FPM setup
• Optimized for Kubernetes deployment
• ReadWriteMany volume support for scaling

MySQL

• Production-ready configuration
• Persistent storage with proper permissions

Scaling

The deployment supports horizontal scaling thanks to ReadWriteMany persistent volumes:

# Scale WordPress pods
kubectl scale deployment my-release-wordpress --replicas=3

# Verify scaling
kubectl get pods -l app=wordpress

Configuration

Edit values.yaml to customize:

• Docker image tags (nginx, wordpress, mysql)
• Replica count (default: 2)
• Storage size and class
• Persistent volume sizes (WordPress: 5Gi, MySQL: 5Gi)

Resources

• Helm Charts – https://helm.sh/docs/
• Prometheus – https://prometheus.io/
• Grafana – https://grafana.com/