DaemonSets
A DaemonSet runs exactly one copy of a Pod on every node (or a selected subset).
Sometimes you need a Pod on every node — not a fixed count, but one per machine. That is what a DaemonSet does.
Typical uses
- Log collectors (like Fluent Bit) that gather logs from each node.
- Monitoring agents (like a node exporter) that report node metrics.
- Networking or storage daemons that must run cluster-wide.
How the count works
You do not set replicas. As nodes join the cluster, the DaemonSet automatically adds a Pod to each; as nodes leave, their Pods are removed. You can restrict it to certain nodes with a nodeSelector or tolerations.
Example
apiVersion: apps/v1
kind: DaemonSet
metadata:
name: node-logger
spec:
selector:
matchLabels:
app: node-logger
template:
metadata:
labels:
app: node-logger
spec:
containers:
- name: fluent-bit
image: fluent/fluent-bit:3.0
volumeMounts:
- name: varlog
mountPath: /var/log
volumes:
- name: varlog
hostPath:
path: /var/logWhen to use it
- A platform team deploys a log collector DaemonSet so every node ships its container logs to a central Elasticsearch cluster automatically.
- A security team runs a vulnerability scanner DaemonSet that checks every node's running containers and reports findings to a dashboard.
- A network team uses a DaemonSet to deploy a CNI plugin agent to every node so pod networking works uniformly across the cluster.
More examples
Log-collector DaemonSet
Deploys a Fluentd log collector to every node, mounting the host /var/log directory to ship node and container logs.
apiVersion: apps/v1
kind: DaemonSet
metadata:
name: fluentd
namespace: kube-system
spec:
selector:
matchLabels:
name: fluentd
template:
metadata:
labels:
name: fluentd
spec:
containers:
- name: fluentd
image: fluentd:v1.16
volumeMounts:
- name: varlog
mountPath: /var/log
volumes:
- name: varlog
hostPath:
path: /var/logVerify DaemonSet pod placement
Confirms the DaemonSet has one desired pod per node, and that all pods are current and ready.
kubectl get ds -n kube-system
# NAME DESIRED CURRENT READY NODE SELECTOR
# fluentd 3 3 3 <none>
kubectl get pods -n kube-system -l name=fluentd -o wide
# One pod per nodeLimit DaemonSet to specific nodes
Restricts DaemonSet pod scheduling to nodes with matching labels, useful for GPU-specific or OS-specific agents.
spec:
template:
spec:
nodeSelector:
kubernetes.io/os: linux
node-role: gpu-worker
containers:
- name: gpu-monitor
image: gpu-monitor:1.0
Discussion