This guide covers cloud-specific configurations for deploying ClickStack on managed Kubernetes services. For basic installation, see the main Helm deployment guide.
Google Kubernetes Engine (GKE)
When deploying to GKE, you may need to override certain values due to cloud-specific networking behavior.
LoadBalancer DNS resolution issue
GKE’s LoadBalancer service can cause internal DNS resolution issues where pod-to-pod communication resolves to external IPs instead of staying within the cluster network. This specifically affects the OTEL collector’s connection to the OpAMP server.
Symptoms:
- OTEL collector logs showing “connection refused” errors with cluster IP addresses
- OpAMP connection failures like:
dial tcp 34.118.227.30:4320: connect: connection refused
Solution:
Use the fully qualified domain name (FQDN) for the OpAMP server URL:
helm install my-clickstack clickstack/clickstack \
--set hyperdx.frontendUrl="http://your-external-ip-or-domain.com" \
--set otel.opampServerUrl="http://my-clickstack-clickstack-app.default.svc.cluster.local:4320"Other GKE considerations
# values-gke.yaml
hyperdx:
frontendUrl: "http://34.123.61.99" # Use your LoadBalancer external IP
otel:
opampServerUrl: "http://my-clickstack-clickstack-app.default.svc.cluster.local:4320"
# Adjust for GKE pod networking if needed
clickhouse:
config:
clusterCidrs:
- "10.8.0.0/16" # GKE commonly uses this range
- "10.0.0.0/8" # Fallback for other configurationsAmazon EKS
For EKS deployments, consider these common configurations:
# values-eks.yaml
hyperdx:
frontendUrl: "http://your-alb-domain.com"
# EKS typically uses these pod CIDRs
clickhouse:
config:
clusterCidrs:
- "192.168.0.0/16"
- "10.0.0.0/8"
# Enable ingress for production
hyperdx:
ingress:
enabled: true
host: "hyperdx.yourdomain.com"
tls:
enabled: trueAzure AKS
For AKS deployments:
# values-aks.yaml
hyperdx:
frontendUrl: "http://your-azure-lb.com"
# AKS pod networking
clickhouse:
config:
clusterCidrs:
- "10.244.0.0/16" # Common AKS pod CIDR
- "10.0.0.0/8"Production Cloud deployment checklist
Before deploying ClickStack to production on any cloud provider:
- Configure proper
frontendUrlwith your external domain/IP - Set up ingress with TLS for HTTPS access
- Override
otel.opampServerUrlwith FQDN if experiencing connection issues (especially on GKE) - Adjust
clickhouse.config.clusterCidrsfor your pod network CIDR - Configure persistent storage for production workloads
- Set appropriate resource requests and limits
- Enable monitoring and alerting
- Configure backup and disaster recovery
- Implement proper secret management
Production best practices
Resource management
hyperdx:
resources:
requests:
cpu: 500m
memory: 1Gi
limits:
cpu: 2000m
memory: 4GiHigh availability
hyperdx:
replicaCount: 3
affinity:
podAntiAffinity:
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 100
podAffinityTerm:
labelSelector:
matchExpressions:
- key: app.kubernetes.io/name
operator: In
values:
- clickstack
topologyKey: kubernetes.io/hostnamePersistent storage
Ensure persistent volumes are configured for data retention:
clickhouse:
persistence:
enabled: true
size: 100Gi
storageClass: "fast-ssd" # Use cloud-specific storage classCloud-specific storage classes:
- GKE:
pd-ssdorpd-balanced - EKS:
gp3orio2 - AKS:
managed-premiumormanaged-csi
Browser compatibility notes
For HTTP-only deployments (development/testing), some browsers may show crypto API errors due to secure context requirements. For production deployments, always use HTTPS with proper TLS certificates through ingress configuration.
See Ingress configuration for TLS setup instructions.
Next steps
- Configuration guide (v1.x) - API keys, secrets, and ingress
- Deployment options (v1.x) - External systems configuration
- Main Helm guide (v1.x) - Basic installation
- Cloud deployments (v2.x) - v2.x cloud guide
- Upgrade guide - Migrating from v1.x to v2.x