Bits of .NET

📡 One Pod Per Node — Automatically

Need monitoring agent on every node? Log collector? DaemonSet runs one pod per node. New nodes get pod automatically.

📝 DaemonSet Example

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: node-monitor
  namespace: kube-system
spec:
  selector:
    matchLabels:
      name: node-monitor
  template:
    metadata:
      labels:
        name: node-monitor
    spec:
      containers:
      - name: monitor
        image: prom/node-exporter
        ports:
        - containerPort: 9100
        resources:
          requests:
            memory: 64Mi
            cpu: 50m
          limits:
            memory: 128Mi
            cpu: 100m

🎯 Common Use Cases

- Node monitoring (Prometheus Node Exporter)
- Log collection (Fluentd, Filebeat)
- Network proxies (kube-proxy)
- Security agents (container scanning)
- CNI plugins (Calico, Cilium)
- Storage drivers (CSI)

# Update DaemonSet
kubectl set image daemonset/node-monitor monitor=prom/node-exporter:v1.3.0

# Rollback
kubectl rollout undo daemonset/node-monitor

💡 Commands

• kubectl get daemonset -n kube-system
• kubectl describe daemonset node-monitor -n kube-system
• kubectl delete daemonset node-monitor -n kube-system
• Use nodeSelector to run on specific nodes