Learning Kubernetes From Scratch (Part 9) – Jobs

🧠 Understanding Kubernetes Jobs

At some point, every system needs to run one-off tasks or scheduled background work — things like backups, data processing, or cleanup scripts.

This is exactly where Kubernetes Jobs and CronJobs come in.

In this guide, we’ll build both — properly, cleanly, and in a way you can actually reuse in real projects.

⚙️ 1. One-Time Job (Pi Calculator)

Let’s start simple. We’ll create a Job that calculates Pi. This isn’t about maths — it’s about understanding how Kubernetes runs finite tasks and tracks their completion.

📄 File: 01-simple-job.yaml

apiVersion: batch/v1
kind: Job

metadata:
  name: pi-calculator

spec:

  # Controls how many retries Kubernetes will attempt if the job fails
  backoffLimit: 4

  template:
    spec:

      containers:
      - name: pi

        # Using a Perl image to calculate Pi
        image: perl:5.34

        # This command calculates Pi to 2000 decimal places
        command:
        - "perl"
        - "-Mbignum=bpi"
        - "-wle"
        - "print bpi(2000)"

      # Important: Job should NOT restart once completed
      restartPolicy: Never

Let’s break down what’s actually happening here 👇

• 📦 Kubernetes creates a pod to run this task
• 🧮 The container executes the Pi calculation
• ✅ Once done, the pod exits and the Job is marked Completed
• 🔁 If it fails, Kubernetes retries (up to 4 times)

🚀 Run the Job

kubectl apply -f 01-simple-job.yaml

👀 Check Job Status

kubectl get jobs

📜 View Output

kubectl logs job/pi-calculator

You should see a massive number — that’s Pi calculated inside your cluster 🎯

⏰ 2. Scheduled Jobs (CronJobs)

Now let’s level up.

A CronJob is just a Job that runs on a schedule — very similar to Linux cron.

📄 File: 02-cronjob.yaml

apiVersion: batch/v1
kind: CronJob

metadata:
  name: backup-job

spec:

  # Cron format: MIN HOUR DAY MONTH DAY-OF-WEEK
  # This runs every day at 2 AM
  schedule: "0 2 * * *"

  jobTemplate:
    spec:
      template:
        spec:

          containers:
          - name: backup

            image: busybox

            # Simulating a backup task with a timestamp
            command:
            - "/bin/sh"
            - "-c"
            - "echo Performing backup at $(date)"

          # Restart ONLY if it fails
          restartPolicy: OnFailure

This is where things get interesting 👇

• ⏰ Kubernetes schedules the job automatically
• 📦 Each run creates a new Job + Pod
• 📊 You can track every execution

🚀 Deploy CronJob

kubectl apply -f 02-cronjob.yaml

👀 Check CronJob

kubectl get cronjob backup-job

🧪 Test It Immediately (Best Practice)

kubectl create job --from=cronjob/backup-job backup-test

📜 View Output

kubectl logs job/backup-test

You’ll see a timestamp printed — simulating a real backup run 🗂️

⚖️ Jobs vs CronJobs

🧹 Cleanup Commands

kubectl delete job pi-calculator
kubectl delete cronjob backup-job
kubectl delete job backup-test

Always clean up after testing — Jobs can pile up quickly and clutter your cluster.

🧠 Final Thoughts

Jobs and CronJobs are one of those features that seem simple… but become critical in real-world systems.

• 🔄 Automating backups
• 📊 Running analytics
• 🧹 Cleaning old data
• ⚙️ Running maintenance scripts

Once you understand this properly, you’ve unlocked a huge part of how real DevOps systems operate behind the scenes.