Tuning node kernel parameters

Kubernetes DaemonSet schedules pod on in each, and we can use it to configure linux kernel parameters.

Using the example below you could configure linux kernel using sysctl depending on your requirements.

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: node-kernel-tuning
spec:
  selector:
    matchLabels:
      app: node-kernel-tuning
  template:
    metadata:
      labels:
        app: node-kernel-tuning
    spec:
      hostNetwork: yes
      initContainers:
        - name: init
          image: alpine:3.14
          command:
            - /bin/sh
            - -xc
            - |
              sysctl net.ipv4.tcp_keepalive_time=7200
              sysctl fs.inotify.max_user_watches=524288
          securityContext:
            privileged: true
      containers:
        - name: sleep
          image: alpine:3.14
          command:
            - /bin/sh
            - -c
            - |
              while true; do sleep 60s; done
          resources:
            requests:
              cpu: 10m
              memory: 10Mi
            limits:
              cpu: 10m
              memory: 10Mi

Using DaemonSet is not ideal as it still requires keep dummy pod running on each node. There is an open issue to support running one time job on each node.