Connect Kubernetes to Digital Tap AI

1 Prerequisites

• Kubernetes cluster — K8s, EKS, GKE, or AKS (v1.24+)
• kubectl access with cluster-admin or equivalent
• Helm 3 installed
• Digital Tap AI account — sign up free

2 RBAC Requirements

The Digital Tap AI agent needs a ClusterRole with these permissions:

CopyapiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: digitaltap-agent
rules:
  # Read cluster state
  - apiGroups: [""]
    resources: ["nodes", "pods", "services", "namespaces", "resourcequotas"]
    verbs: ["get", "list", "watch"]
  - apiGroups: [""]
    resources: ["pods/log"]
    verbs: ["get"]
  # Read metrics
  - apiGroups: ["metrics.k8s.io"]
    resources: ["nodes", "pods"]
    verbs: ["get", "list"]
  # Read deployments, statefulsets, daemonsets
  - apiGroups: ["apps"]
    resources: ["deployments", "statefulsets", "daemonsets", "replicasets"]
    verbs: ["get", "list", "watch"]
  # Optimization actions (remove for monitor-only)
  - apiGroups: ["apps"]
    resources: ["deployments", "statefulsets"]
    verbs: ["patch", "update"]
  - apiGroups: ["autoscaling"]
    resources: ["horizontalpodautoscalers"]
    verbs: ["get", "list", "watch", "patch", "update"]
  # Read events for anomaly detection
  - apiGroups: [""]
    resources: ["events"]
    verbs: ["get", "list", "watch"]

3 Install via Helm (Recommended)

Copy# Add the Digital Tap Helm repo
helm repo add digitaltap https://charts.digitaltap.ai
helm repo update

# Install the agent
helm install digitaltap-agent digitaltap/agent \
  --set apiKey="your-digital-tap-api-key" \
  --set platform="kubernetes" \
  --set clusterName="my-production-cluster" \
  --namespace digitaltap \
  --create-namespace

Custom values.yaml

Copy# values.yaml
apiKey: "your-digital-tap-api-key"
platform: kubernetes
clusterName: my-production-cluster

# Scan interval (seconds)
scanInterval: 180

# Start in dry-run mode
dryRun: true

# Resource limits for the agent itself
resources:
  requests:
    cpu: 100m
    memory: 128Mi
  limits:
    cpu: 500m
    memory: 256Mi

# Namespaces to monitor (empty = all)
namespaces: []

# Namespaces to exclude
excludeNamespaces:
  - kube-system
  - kube-public

Copyhelm install digitaltap-agent digitaltap/agent \
  -f values.yaml \
  --namespace digitaltap --create-namespace

4 Verify Installation

Copy# Check agent is running
kubectl get pods -n digitaltap

# Check agent logs
kubectl logs -n digitaltap -l app=digitaltap-agent --tail=50

# Expected output:
# [INFO] Connected to Digital Tap AI
# [INFO] Discovered 42 pods across 8 namespaces
# [INFO] First scan complete — 12 optimization opportunities found

5 K8s-Specific Agents & Features

• Pod Right-Sizing — Analyzes actual CPU/memory usage vs requests/limits and recommends optimal values
• Node Optimization — Identifies under-utilized nodes and recommends consolidation or instance type changes
• Bin Packing — Improves pod placement to maximize node utilization and reduce node count
• HPA Tuning — Optimizes HorizontalPodAutoscaler targets based on actual scaling patterns
• Spot Node Detection — Identifies workloads safe for spot/preemptible nodes and recommends migration
• Idle Namespace Detection — Finds namespaces with low or zero resource utilization
• Resource Quota Optimization — Recommends namespace quota adjustments based on actual usage

Namespace-Level Cost Attribution

Digital Tap AI automatically breaks down costs by namespace, providing:

• Per-namespace compute cost (CPU + memory)
• Per-namespace storage cost (PV/PVC)
• Per-namespace network cost (where available)
• Team-level rollup via namespace labels
• Daily/weekly/monthly trend analysis

6 Troubleshooting

Agent CrashLoopBackOff

• Check logs: kubectl logs -n digitaltap -l app=digitaltap-agent
• Verify API key is correct
• Ensure RBAC ClusterRole and ClusterRoleBinding are applied

Metrics unavailable

• Ensure metrics-server is installed: kubectl top nodes
• For EKS, install metrics-server: kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml