
For a project I needed a way to automatically schedule pods to be deleted every 3 hours. This was due to the nature of the project where we wanted a cleanup in memory due to zombie processes appearing. It's not a clean fix, but it was necessary for production to run stable while we were patching things.

Setting this up is quite easy actually! Just create the script below and save it to a file of your preferences (e.g. `deploy-cronjob.yaml`), finally, applying it through `kubectl apply -f deploy-cronjob.yaml`.

What this YAML does:
1. It creates the necessary Service Accounts and Roles and Role Bindings to have the permissions to let an automated job run `kubectl delete` 
2. It spins up a `bitnami/kubectl` container
3. It runs the `/bin/sh -c <your_command>` that will get executed

> ⚠️ We utilize `/bin/sh -c` since we utilize command substitution which is only available in the Shell.

**deploy-cronjob.yaml**

```yaml
# Apply With:
# kubectl apply -f deploy-cronjob.yaml

# Validate with:
# kubectl get cronjob

---
kind: ServiceAccount
apiVersion: v1
metadata:
  name: deleting-pods
  namespace: default

---
apiVersion: rbac.authorization.k8s.io/v1
kind: Role
metadata:
  name: deleting-pods
  namespace: default
rules:
  - apiGroups: [""]
    resources: ["pods"]
    verbs: ["get", "patch", "list", "watch", "delete"]

---
apiVersion: rbac.authorization.k8s.io/v1
kind: RoleBinding
metadata:
  name: deleting-pods
  namespace: default
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: Role
  name: deleting-pods
subjects:
  - kind: ServiceAccount
    name: deleting-pods

---
apiVersion: batch/v1beta1
kind: CronJob
metadata:
  name: deleting-pod-myname
  namespace: default
spec:
  concurrencyPolicy: Forbid
  schedule: "0 */3 * * *" # At minute 0 past every 6 hours
  jobTemplate:
    spec:
      backoffLimit: 2
      activeDeadlineSeconds: 600
      template:
        spec:
          serviceAccountName: deleting-pods
          restartPolicy: Never
          containers:
            - name: kubectl
              image: bitnami/kubectl
              command: [ "/bin/sh", "-c" ]
              args: 
                - 'kubectl delete pod $(kubectl get pod -l app=<your_label> -o jsonpath="{.items[0].metadata.name}")'
```

Learn how to schedule a cronjob in Kubernetes that will delete pods automatically and restart it through the deployment