Deploy on Azure AKS

This guide covers deploying the Hiya Voice Verification engine on Azure Kubernetes Service (AKS). It builds on the generic Kubernetes guide with AKS-specific configuration.

Prerequisites

• An AKS cluster (v1.24+) with kubectl configured
• Container image pulled and authenticated — see Getting the Container Image
• Runtime configuration values for API_KEY, ORG_HANDLE, PLATFORM_REGION, and MIN_ALLOCATION
• Azure CLI (az) installed

Recommended VM Series

We recommend Intel Emerald Rapids-based VMs for optimal performance. The following Azure v6 VM series run on Intel Xeon Platinum 8573C (Emerald Rapids) at 3.0 GHz all-core turbo:

VM Series	Category	Notes
Dsv6 / Ddsv6	General-purpose	High vCPU counts available
Dlsv6 / Dldsv6	General-purpose (low memory)	High vCPU counts available
Esv6 / Edsv6	Memory-optimized	Memory-optimized series with large RAM capacity
Ebsv6 / Ebdsv6	Memory-optimized (storage bandwidth)	Memory-optimized with high storage bandwidth
FXmsv2 / FXmdsv2	Compute-optimized	4.0 GHz all-core turbo
Lsv4	Storage-optimized	High-throughput local NVMe

Azure also offers AMD-based v6 families (e.g., Dasv6, Easv6). These are supported but Intel Emerald Rapids is recommended for best performance.

Ensure your node pool VMs have sufficient RAM per pod to hold the engine and ML models in memory. Contact Hiya for sizing guidance.

Step 1 — Create Secrets

Create the image pull secret and runtime configuration secret:

kubectl create secret docker-registry hiya-registry \
  --docker-server=europe-docker.pkg.dev \
  --docker-username=_json_key \
  --docker-password="$(cat key.json)"

kubectl create secret generic hiya-engine-config \
  --from-literal=api-key=<your-api-key> \
  --from-literal=org-handle=<your-org-handle> \
  --from-literal=platform-region=<eu-or-us> \
  --from-literal=min-allocation=1m

Alternative — Azure Container Registry Mirror

If you prefer to mirror the image in your own Azure Container Registry:

# Import directly from Google Artifact Registry
az acr import \
  --name <your-acr-name> \
  --source europe-docker.pkg.dev/loccus-platform/onpremise-images/hiya-voice-verification:<version> \
  --image hiya/hiya-voice-verification:<version> \
  --username _json_key \
  --password "$(cat key.json)"

Then reference <your-acr-name>.azurecr.io/hiya/hiya-voice-verification:<version> in your deployment manifest and attach the ACR to your AKS cluster:

az aks update \
  --name <your-aks-cluster> \
  --resource-group <your-resource-group> \
  --attach-acr <your-acr-name>

Step 2 — Deploy

Apply the Deployment and Service manifests from the Kubernetes guide. No AKS-specific changes are needed.

Step 3 — Expose via Internal Load Balancer (Optional)

For clients outside the cluster, use an Azure internal load balancer:

# hiya-ilb-service.yaml
apiVersion: v1
kind: Service
metadata:
  name: hiya-voice-verification-ilb
  annotations:
    service.beta.kubernetes.io/azure-load-balancer-internal: "true"
spec:
  type: LoadBalancer
  selector:
    app: hiya-voice-verification
  ports:
    - name: health
      protocol: TCP
      port: 8080
      targetPort: 8080
    - name: ws
      protocol: TCP
      port: 8081
      targetPort: 8081

Network Configuration

Ensure the following outbound access from your AKS cluster:

Destination	Port	Protocol	Purpose
europe-docker.pkg.dev	443	HTTPS	Image pulls
api.hiya.com	443	HTTPS	License verification and billing

For private AKS clusters, ensure your Azure Firewall or NSG rules allow outbound HTTPS to these destinations.

Scaling

Use the Horizontal Pod Autoscaler for automatic scaling:

kubectl autoscale deployment hiya-voice-verification \
  --cpu-percent=50 \
  --min=1 \
  --max=10

For node-level scaling, enable the AKS cluster autoscaler on your node pool:

az aks nodepool update \
  --resource-group <your-resource-group> \
  --cluster-name <your-aks-cluster> \
  --name <your-nodepool> \
  --enable-cluster-autoscaler \
  --min-count 1 \
  --max-count 5