Skip to main content

Kubernetes Deployment

Deploy Ovyxa on Kubernetes for high availability, auto-scaling, and enterprise-grade reliability.

Prerequisites​

  • Kubernetes cluster 1.25+ (EKS, GKE, AKS, or self-managed)
  • kubectl configured
  • Helm 3.x installed
  • 3+ nodes recommended (high availability)
  • Persistent volume provisioner (gp3, pd-ssd, etc.)

Quick Install with Helm​

1. Add Helm Repository​

helm repo add ovyxa https://charts.ovyxa.com
helm repo update

2. Create Namespace​

kubectl create namespace ovyxa

3. Install Chart​

helm install ovyxa ovyxa/ovyxa \
  --namespace ovyxa \
  --set ingress.hosts[0].host=analytics.yourdomain.com \
  --set ingress.tls[0].secretName=ovyxa-tls \
  --set ingress.tls[0].hosts[0]=analytics.yourdomain.com \
  --set postgresql.auth.password=$(openssl rand -hex 32) \
  --set redis.auth.password=$(openssl rand -hex 32)

4. Wait for Pods​

kubectl wait --for=condition=ready pod \
  -l app.kubernetes.io/instance=ovyxa \
  -n ovyxa \
  --timeout=300s

5. Access Dashboard​

# Get external IP
kubectl get ingress -n ovyxa

# Visit https://analytics.yourdomain.com

Custom values.yaml​

Create values.yaml for advanced configuration:

# Global settings
global:
  domain: analytics.yourdomain.com

# Ingress
ingress:
  enabled: true
  className: nginx
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-prod
  hosts:
    - host: analytics.yourdomain.com
      paths:
        - path: /
          pathType: Prefix
  tls:
    - secretName: ovyxa-tls
      hosts:
        - analytics.yourdomain.com

# API (ingestion + stats)
api:
  replicaCount: 3
  image:
    repository: ovyxa/api
    tag: latest
    pullPolicy: IfNotPresent
  resources:
    requests:
      cpu: 500m
      memory: 512Mi
    limits:
      cpu: 2000m
      memory: 2Gi
  autoscaling:
    enabled: true
    minReplicas: 3
    maxReplicas: 10
    targetCPUUtilizationPercentage: 70
  env:
    - name: NODE_ENV
      value: production
    - name: LOG_LEVEL
      value: info

# Web dashboard
web:
  replicaCount: 2
  image:
    repository: ovyxa/web
    tag: latest
  resources:
    requests:
      cpu: 200m
      memory: 256Mi
    limits:
      cpu: 1000m
      memory: 512Mi

# PostgreSQL (metadata)
postgresql:
  enabled: true
  auth:
    database: ovyxa
    username: ovyxa
    password: ""  # Auto-generated if empty
  primary:
    persistence:
      enabled: true
      size: 20Gi
      storageClass: gp3  # AWS example
    resources:
      requests:
        cpu: 500m
        memory: 1Gi
      limits:
        cpu: 2000m
        memory: 4Gi

# ClickHouse (analytics)
clickhouse:
  enabled: true
  replicaCount: 3  # For HA
  persistence:
    enabled: true
    size: 500Gi
    storageClass: gp3
  resources:
    requests:
      cpu: 2000m
      memory: 4Gi
    limits:
      cpu: 8000m
      memory: 16Gi
  zookeeper:
    enabled: true  # Required for replication
    replicaCount: 3

# Redis (cache)
redis:
  enabled: true
  architecture: standalone  # or 'replication' for HA
  auth:
    enabled: true
    password: ""  # Auto-generated
  master:
    resources:
      requests:
        cpu: 200m
        memory: 256Mi
      limits:
        cpu: 1000m
        memory: 1Gi
    persistence:
      enabled: true
      size: 10Gi

Install with custom values:

helm install ovyxa ovyxa/ovyxa \
  --namespace ovyxa \
  --values values.yaml

High Availability Setup​

For production HA:

# values-ha.yaml
api:
  replicaCount: 5
  podAntiAffinity: hard  # Spread across nodes

web:
  replicaCount: 3
  podAntiAffinity: soft

postgresql:
  architecture: replication
  replicaCount: 3

clickhouse:
  replicaCount: 3
  shards: 2  # Horizontal sharding

redis:
  architecture: replication
  sentinel:
    enabled: true

Monitoring​

Prometheus + Grafana​

Enable metrics:

# values.yaml
metrics:
  enabled: true
  serviceMonitor:
    enabled: true
    namespace: monitoring

Install Prometheus Operator:

helm install prometheus prometheus-community/kube-prometheus-stack \
  --namespace monitoring \
  --create-namespace

Access Grafana:

kubectl port-forward -n monitoring svc/prometheus-grafana 3000:80
# Visit http://localhost:3000
# Default: admin/prom-operator

Import Ovyxa dashboard (ID: 12345) from Grafana.com.

Scaling​

Manual Scaling​

# Scale API pods
kubectl scale deployment ovyxa-api \
  --replicas=10 \
  -n ovyxa

# Scale ClickHouse
kubectl scale statefulset ovyxa-clickhouse \
  --replicas=5 \
  -n ovyxa

Auto-scaling (HPA)​

Already configured in Helm chart. Monitor:

kubectl get hpa -n ovyxa

Backup & Restore​

Using Velero​

Install Velero:

velero install \
  --provider aws \
  --plugins velero/velero-plugin-for-aws:v1.8.0 \
  --bucket ovyxa-backups \
  --backup-location-config region=us-east-1 \
  --snapshot-location-config region=us-east-1

Schedule backups:

velero schedule create ovyxa-daily \
  --schedule="0 2 * * *" \
  --include-namespaces ovyxa \
  --ttl 720h0m0s

Restore:

velero restore create --from-backup ovyxa-daily-20250113

Upgrade​

# Update Helm repo
helm repo update

# Upgrade release
helm upgrade ovyxa ovyxa/ovyxa \
  --namespace ovyxa \
  --values values.yaml \
  --reuse-values

# Check status
helm status ovyxa -n ovyxa

Troubleshooting​

Pods Not Starting​

# Check pod status
kubectl get pods -n ovyxa

# Describe problematic pod
kubectl describe pod ovyxa-api-xxx -n ovyxa

# View logs
kubectl logs -f ovyxa-api-xxx -n ovyxa

PVC Not Binding​

Check storage class:

kubectl get storageclass
kubectl get pvc -n ovyxa

Ensure provisioner is configured correctly.

High Memory Usage​

Check ClickHouse queries:

kubectl exec -it ovyxa-clickhouse-0 -n ovyxa -- \
  clickhouse-client --query "SELECT * FROM system.query_log ORDER BY memory_usage DESC LIMIT 10"

Cost Optimization​

AWS EKS Example​

# Use spot instances for non-critical workloads
web:
  nodeSelector:
    kubernetes.io/lifecycle: spot

# Use reserved instances for databases
clickhouse:
  nodeSelector:
    kubernetes.io/lifecycle: on-demand

Right-sizing​

Monitor actual usage:

kubectl top pods -n ovyxa

Adjust requests/limits accordingly.

Next Steps​

Kubernetes is ideal for 100M+ events/month and enterprise requirements.