01_check_pod_health.yml - Usage Guide

Read-Only k3s kubeadm

Purpose

This playbook reports the current status and health of pods in all namespaces of your Kubernetes cluster (typically run on master or singlenode): whether they are healthy, restart counts, the node they run on, and so on.

Requirements

  • kubectl must be installed and configured on the target host (master or singlenode).
  • Your Ansible inventory must define a [master] or [singlenode] group.
  • The playbook explicitly sets KUBECONFIG to ~/.kube/config because .bashrc is not loaded in non-interactive SSH (especially on k3s). Classic kubeadm installs use the same default path.

The playbook now runs a kubectl cluster-info access check before collecting pod data. If kubectl is unavailable (missing/wrong KUBECONFIG or the cluster is unreachable), a clear warning is printed and the pod query is skipped.

How to run 

# Run for customer A
ansible-playbook -i inventories/musteri_a/hosts.ini playbooks/01_check_pod_health.yml

# Run for customer B
ansible-playbook -i inventories/musteri_b/hosts.ini playbooks/01_check_pod_health.yml

Sample output

When the command runs, a ping connectivity test runs first. If it succeeds, the output of kubectl get pods -A -o wide is printed:

TASK [Ping connectivity test] **************************************************
ok: [192.168.1.10]

TASK [Check kubeconfig / kubectl access] ***************************************
ok: [192.168.1.10]

TASK [Get all pods in all namespaces with wide output] *************************
ok: [192.168.1.10]

TASK [Display pod health status] ***********************************************
ok: [192.168.1.10] => {
    "msg": [
        "NAMESPACE     NAME                               READY   STATUS    RESTARTS   AGE   IP            NODE      NOMINATED NODE   READINESS GATES",
        "kube-system   calico-node-abcd1                  1/1     Running   0          12d   192.168.1.10  master    <none>           <none>",
        "argocd        argocd-server-54d68c4847-p9qwx     1/1     Running   0          5d    10.244.1.5    worker1   <none>           <none>"
    ]
}