Why should you care?

Hospital peak hours and CRM Monday spikes need HPA on agent workers without manual VM scaling.

See it live — copy this example

Paste into an ASP.NET Core 8+ / AgentNest project, then run with dotnet run (set your API keys in user-secrets).

# agentnest-api-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: agentnest-api
spec:
  replicas: 3
  template:
    spec:
      containers:
      - name: api
        image: agentnest/api:1.4.0
        envFrom:
        - secretRef:
            name: azure-openai
        readinessProbe:
          httpGet: { path: /health/ready, port: 8080 }

What happened?

• Deployment runs three API replicas with secrets for Azure OpenAI and readiness gate before receiving traffic.
• Follow the steps below — typing the code yourself is the fastest way to learn.

Practice next

1. Create AKS cluster with workload identity.
2. kubectl apply deployment and ClusterIP service.
3. Configure HPA on CPU and custom metric ai_queue_depth.
4. Add PodDisruptionBudget minAvailable 2.
5. Schedule GPU embedding jobs on separate node pool.

Remember

Run AgentNest API on Kubernetes with probes and secrets. Scale workers separately from stateless API pods. Use rolling updates for agent platform releases.

SaaS peak scaling

Black Friday CRM copilot traffic triples concurrent streams.

Outcome: HPA adds API pods; KEDA scales embedding workers from Service Bus depth.