Build a Node-Level Log Collector DaemonSet

The situation

There's no per-node log collection. You need an agent that runs one pod on every node in the cluster — and automatically on any node that joins later. That workload shape is a DaemonSet, not a Deployment.

                                           ┌─────────┐
                           ┌─────1 pod────▶│worker-1 │
 ┌──────────────────────┐──┘               │         │
 │DaemonSet node-logger │                  └─────────┘
 │                      │──┐               ┌─────────┐
 └──────────────────────┘  └─────1 pod────▶│worker-2 │
                                           │         │
                                           └─────────┘

Your task

In the kubelings namespace, create a DaemonSet named node-logger:

One pod per node (DaemonSets do this by design — no replicas).
Use image ghcr.io/iximiuz/labs/busybox:latest, command e.g. sh -c 'while true; do echo collecting; sleep 3600; done'.
All its pods must reach Ready.

kubectl -n kubelings get ds,pods -o wide

Hint

kubectl -n kubelings create -f - <<'EOF'
apiVersion: apps/v1
kind: DaemonSet
metadata: {name: node-logger}
spec:
  selector: {matchLabels: {app: node-logger}}
  template:
    metadata: {labels: {app: node-logger}}
    spec:
      containers:
        - name: agent
          image: ghcr.io/iximiuz/labs/busybox:latest
          command: ["sh","-c","while true; do echo collecting; sleep 3600; done"]
EOF

Control-plane nodes are schedulable in this playground, so no tolerations are strictly required — but adding a control-plane toleration is good practice.

Solution

Approach

A per-node agent is a DaemonSet — it schedules exactly one pod per eligible node and tracks node add/remove automatically.

Create it

kubectl -n kubelings apply -f - <<'EOF'
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: node-logger
spec:
  selector:
    matchLabels: {app: node-logger}
  template:
    metadata:
      labels: {app: node-logger}
    spec:
      tolerations:
        - key: node-role.kubernetes.io/control-plane
          operator: Exists
          effect: NoSchedule
      containers:
        - name: agent
          image: ghcr.io/iximiuz/labs/busybox:latest
          command: ["sh","-c","while true; do echo collecting; sleep 3600; done"]
EOF

Verify

kubectl -n kubelings get ds node-logger
kubectl -n kubelings get pods -l app=node-logger -o wide

DESIRED should equal READY and match the node count.