{"id":20595,"date":"2026-07-04T05:02:37","date_gmt":"2026-07-04T05:02:37","guid":{"rendered":"http:\/\/www.devopsschool.com\/blog\/?p=20595"},"modified":"2026-07-04T05:02:41","modified_gmt":"2026-07-04T05:02:41","slug":"explained-kubernetes-jobs-and-cronjobs-with-example","status":"publish","type":"post","link":"https:\/\/www.devopsschool.com\/blog\/explained-kubernetes-jobs-and-cronjobs-with-example\/","title":{"rendered":"Kubernetes Job & CronJob Explained: API Resource, YAML Example, and Use Cases"},"content":{"rendered":"\n

<\/p>\n\n\n\n

Kubernetes Jobs and CronJobs Explained with Examples: Complete Hands-on Guide for DevOps Engineer<\/strong><\/p>\n\n\n\n

Keywords<\/h2>\n\n\n
Kubernetes Jobs\nKubernetes CronJobs\nKubernetes Job example\nKubernetes CronJob example\nKubernetes batch workloads\nkubectl create job\nkubectl create cronjob\nKubernetes scheduled jobs\nKubernetes background tasks\nKubernetes batch processing\n<\/code><\/span><\/pre>\n\n\n
\n\n\n\n
1. Introduction<\/h1>\n\n\n\n
In Kubernetes, not every workload is a long-running application like a web server, API service, or microservice. Many real-world tasks are short-lived. They start, perform some work, complete, and then stop.<\/p>\n\n\n\n
Examples include:<\/p>\n\n\n
Database backup\nReport generation\nData cleanup\nFile processing\nBatch import<\/span>\/export<\/span>\nEmail notification batch\nLog archival\nCertificate renewal\nOne-time migration\nScheduled health check\nMachine learning batch job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
For these use cases, Kubernetes provides two important workload resources:<\/p>\n\n\n
Job\nCronJob\n<\/code><\/span><\/pre>\n\n\n
A Job<\/strong> runs one or more Pods until the task completes successfully. A CronJob<\/strong> creates Jobs repeatedly according to a time-based schedule. Kubernetes official documentation defines Jobs as one-off tasks that run to completion, while CronJobs create Jobs on a repeating schedule.<\/p>\n\n\n\n
As of the current Kubernetes documentation baseline, Kubernetes v1.36.2<\/strong> is the latest stable patch release in the Kubernetes 1.36 line, released on June 9, 2026<\/strong>. The examples in this tutorial use the stable batch\/v1<\/code> APIs for Jobs and CronJobs.<\/p>\n\n\n\n
\n\n\n\n
2. Job vs CronJob \u2014 Quick Comparison<\/h1>\n\n\n\n
Feature<\/th>Job<\/th>CronJob<\/th><\/tr><\/thead>
Purpose<\/td>Run a task once until completion<\/td>Run tasks repeatedly on a schedule<\/td><\/tr>
API version<\/td>batch\/v1<\/code><\/td>batch\/v1<\/code><\/td><\/tr>
Creates Pods directly?<\/td>Yes, Job controller creates Pods<\/td>No, CronJob creates Jobs, then Jobs create Pods<\/td><\/tr>
Common use<\/td>One-time migration, batch processing, data import<\/td>Backups, reports, cleanup, scheduled sync<\/td><\/tr>
Scheduling<\/td>Runs immediately after creation unless suspended<\/td>Runs based on cron schedule<\/td><\/tr>
Retry support<\/td>Yes, using backoffLimit<\/code><\/td>Yes, through Job template<\/td><\/tr>
Parallel execution<\/td>Yes, using parallelism<\/code> and completions<\/code><\/td>Yes, inside jobTemplate<\/code><\/td><\/tr>
Automatic cleanup<\/td>Yes, using ttlSecondsAfterFinished<\/code><\/td>Uses Job history limits and\/or Job TTL<\/td><\/tr>
Time zone support<\/td>Not applicable<\/td>Yes, using .spec.timeZone<\/code><\/td><\/tr>
Concurrency control<\/td>Controlled by Job parallelism<\/td>Controlled by concurrencyPolicy<\/code><\/td><\/tr>
Best for<\/td>One-time run-to-completion workloads<\/td>Scheduled run-to-completion workloads<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Simple rule:<\/p>\n\n\n
Use<\/span> Job<\/span> when<\/span> the<\/span> task<\/span> must<\/span> run<\/span> once<\/span>.\nUse<\/span> CronJob<\/span> when<\/span> the<\/span> task<\/span> must<\/span> run<\/span> repeatedly<\/span> on<\/span> a<\/span> schedule<\/span>.\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
3. How Kubernetes Job Works<\/h1>\n\n\n\n
A Kubernetes Job creates one or more Pods and tracks successful completions. The Job is considered complete when the required number of Pods finish successfully. For a simple non-parallel Job, Kubernetes usually starts one Pod, and the Job completes when that Pod exits successfully.<\/p>\n\n\n\n
Basic flow:<\/p>\n\n\n
User creates Job\n        |\n        v\nJob controller creates Pod\n        |\n        v\nContainer runs command\n        |\n        v\nCommand exits with<\/span> code 0<\/span>\n        |\n        v\nPod becomes Completed\n        |\n        v\nJob becomes Complete\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
If the container exits with a non-zero code, Kubernetes can retry according to the Job configuration. By default, the Job uses .spec.backoffLimit<\/code>, which defaults to 6<\/code>, unless per-index backoff is configured for Indexed Jobs.<\/p>\n\n\n\n
\n\n\n\n
4. How Kubernetes CronJob Works<\/h1>\n\n\n\n
A CronJob is like a Linux crontab entry inside Kubernetes. It does not run the container directly. Instead, it creates a Job at the scheduled time, and that Job creates the Pod. Kubernetes documentation defines CronJob as a resource that starts one-time Jobs on a repeating schedule.<\/p>\n\n\n\n
Basic flow:<\/p>\n\n\n
User creates CronJob\n        |\n        v\nCronJob controller watches schedule\n        |\n        v\nAt scheduled time, CronJob creates Job\n        |\n        v\nJob creates Pod\n        |\n        v\nContainer runs command\n        |\n        v\nPod completes\n        |\n        v\nJob completes\n<\/code><\/span><\/pre>\n\n\n
Example:<\/p>\n\n\n
CronJob schedule: every day at 2<\/span> AM\n        |\n        v\nCreates a new<\/span> Job daily at 2<\/span> AM\n        |\n        v\nEach Job runs backup script\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
5. Prerequisites<\/h1>\n\n\n\n
You need access to a Kubernetes cluster.<\/p>\n\n\n\n
Check client and server version:<\/p>\n\n\n
kubectl version\n<\/code><\/span><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc version\n<\/code><\/span><\/pre>\n\n\n
Create a namespace for practice:<\/p>\n\n\n
kubectl create namespace jobs-demo\nkubectl config set<\/span>-context --current --namespace=jobs-demo\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc new<\/span>-project jobs-demo\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Verify:<\/p>\n\n\n
kubectl get<\/span> ns jobs-demo\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Or:<\/p>\n\n\n
oc project\n<\/code><\/span><\/pre>\n\n\n
\n\n\n\n
6. Basic Kubernetes Job Example<\/h1>\n\n\n\n
Let us create a simple Job that prints a message and exits.<\/p>\n\n\n\n
Create file:<\/p>\n\n\n
cat > job-hello.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: hello-job\nspec:\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: hello\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Hello from Kubernetes Job\"<\/span>\n              echo<\/span> \"Current date is:\"<\/span>\n              date\n              echo<\/span> \"Job completed successfully\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-hello<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc<\/span> apply<\/span> -f<\/span> job-hello<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check Job:<\/p>\n\n\n
kubectl get<\/span> jobs\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Expected:<\/p>\n\n\n
NAME        STATUS     COMPLETIONS   DURATION   AGE\nhello-job   Complete   1\/1           5s         10s\n<\/code><\/span><\/pre>\n\n\n
Check Pods:<\/p>\n\n\n
kubectl get<\/span> pods\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Expected:<\/p>\n\n\n
NAME              READY   STATUS      RESTARTS   AGE\nhello-job-xxxxx   0\/1     Completed   0          20s\n<\/code><\/span><\/pre>\n\n\n
View logs:<\/p>\n\n\n
kubectl logs job\/hello-job\n<\/code><\/span><\/pre>\n\n\n
Expected output:<\/p>\n\n\n
Hello<\/span> from<\/span> Kubernetes<\/span> Job<\/span>\nCurrent<\/span> date<\/span> is<\/span>:\nSat<\/span> Jul<\/span>  4 10:00<\/span>:00<\/span> UTC<\/span> 2026\nJob<\/span> completed<\/span> successfully<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
The Pod remains after completion so that you can inspect status and logs. Kubernetes documentation notes that completed Job Pods are usually not deleted immediately, allowing you to inspect logs and diagnostics.<\/p>\n\n\n\n
\n\n\n\n
7. Important Job Fields<\/h1>\n\n\n\n
Field<\/th>Meaning<\/th><\/tr><\/thead>
template<\/code><\/td>Pod template used by the Job<\/td><\/tr>
restartPolicy<\/code><\/td>Must be Never<\/code> or OnFailure<\/code><\/td><\/tr>
backoffLimit<\/code><\/td>Number of retries before Job is marked failed<\/td><\/tr>
activeDeadlineSeconds<\/code><\/td>Maximum time allowed for the whole Job<\/td><\/tr>
completions<\/code><\/td>Number of successful Pods required<\/td><\/tr>
parallelism<\/code><\/td>Maximum Pods running at the same time<\/td><\/tr>
completionMode<\/code><\/td>NonIndexed<\/code> or Indexed<\/code><\/td><\/tr>
ttlSecondsAfterFinished<\/code><\/td>Auto-cleanup time after Job finishes<\/td><\/tr>
suspend<\/code><\/td>Pause Job execution<\/td><\/tr>
podFailurePolicy<\/code><\/td>Advanced failure handling<\/td><\/tr>
successPolicy<\/code><\/td>Advanced success handling for Indexed Jobs<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
The Kubernetes Job API states that restartPolicy<\/code> inside the Job Pod template can only be Never<\/code> or OnFailure<\/code>. It also defines important fields such as activeDeadlineSeconds<\/code>, backoffLimit<\/code>, completionMode<\/code>, parallelism<\/code>, suspend<\/code>, and ttlSecondsAfterFinished<\/code>.<\/p>\n\n\n\n
\n\n\n\n
8. Job Restart Policy: Never<\/code> vs OnFailure<\/code><\/h1>\n\n\n\n
A Job Pod can use:<\/p>\n\n\n
restartPolicy<\/span>: Never\nrestartPolicy<\/span>: OnFailure\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
restartPolicy: Never<\/code><\/h2>\n\n\n\n
If the container fails, the Pod is marked failed and the Job controller creates a new Pod.<\/p>\n\n\n\n
restartPolicy: OnFailure<\/code><\/h2>\n\n\n\n
If the container fails, Kubernetes restarts the container inside the same Pod.<\/p>\n\n\n\n
Kubernetes documentation explains that when a container fails and restartPolicy<\/code> is OnFailure<\/code>, the Pod stays on the node and the container is rerun. If the Pod fails entirely, the Job controller starts a new Pod.<\/p>\n\n\n\n
For clean learning, use:<\/p>\n\n\n
restartPolicy<\/span>: Never\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
For production, choose based on how your application handles retries.<\/p>\n\n\n\n
\n\n\n\n
9. Failed Job Example<\/h1>\n\n\n\n
Create a Job that always fails.<\/p>\n\n\n
cat > job-fail.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: fail-job\nspec:\n  backoffLimit: 3<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: fail\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"This job will fail\"<\/span>\n              exit<\/span> 1<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-fail<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> jobs,pods -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
You will see multiple failed Pods because backoffLimit: 3<\/code> allows retries before the Job is marked failed.<\/p>\n\n\n\n
Check Job details:<\/p>\n\n\n
kubectl describe job fail-job\n<\/code><\/span><\/pre>\n\n\n
Check failed Pods:<\/p>\n\n\n
kubectl get<\/span> pods -l job-name=fail-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check logs from one failed Pod:<\/p>\n\n\n
kubectl logs -l job-name=fail-job\n<\/code><\/span><\/pre>\n\n\n
Expected:<\/p>\n\n\n
This job will fail\n<\/code><\/span><\/pre>\n\n\n
Clean up:<\/p>\n\n\n
kubectl delete<\/span> job fail-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
10. backoffLimit<\/code> \u2014 Retry Control<\/h1>\n\n\n\n
backoffLimit<\/code> controls how many failures are allowed before Kubernetes marks the Job as failed. The default is 6<\/code>, unless backoffLimitPerIndex<\/code> is configured for an Indexed Job.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
spec:\n  backoffLimit: 3\n<\/code><\/span><\/pre>\n\n\n
Meaning:<\/p>\n\n\n
Kubernetes retries failed Pods.\nAfter retry limit is reached, Job becomes Failed.\n<\/code><\/span><\/pre>\n\n\n
Use low values when:<\/p>\n\n\n
Failure means a real bug\nYou do<\/span> not want wasteful retries\nThe task is expensive\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Use higher values when:<\/p>\n\n\n
Failures may be temporary\nNetwork dependencies are unstable\nExternal API may recover\n<\/code><\/span><\/pre>\n\n\n
\n\n\n\n
11. activeDeadlineSeconds<\/code> \u2014 Maximum Runtime<\/h1>\n\n\n\n
activeDeadlineSeconds<\/code> limits the total runtime of the Job. Once the Job exceeds this duration, Kubernetes terminates running Pods and marks the Job as failed with reason DeadlineExceeded<\/code>.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
cat > job-deadline.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: deadline-job\nspec:\n  activeDeadlineSeconds: 20<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: slow-task\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Starting long task\"<\/span>\n              sleep 60<\/span>\n              echo<\/span> \"This line may never run\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-deadline<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> job deadline-job -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Describe:<\/p>\n\n\n
kubectl describe job deadline-job\n<\/code><\/span><\/pre>\n\n\n
Expected reason:<\/p>\n\n\n
DeadlineExceeded\n<\/code><\/span><\/pre>\n\n\n
Clean up:<\/p>\n\n\n
kubectl delete<\/span> job deadline-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
12. Auto-Cleanup Finished Jobs with TTL<\/h1>\n\n\n\n
By default, finished Jobs and their Pods may stay in the cluster. For production, you should clean them up.<\/p>\n\n\n\n
Kubernetes provides ttlSecondsAfterFinished<\/code> for automatic cleanup of completed or failed Jobs. The TTL-after-finished controller is stable and removes the Job and its dependent objects after the configured TTL.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
cat > job-ttl.yaml <<'EOF'<\/span>\napiVersion<\/span>: batch\/v1\nkind<\/span>: Job\nmetadata<\/span>:\n  name: ttl-job\nspec<\/span>:\n  ttlSecondsAfterFinished: 60<\/span>\n  template<\/span>:\n    spec:\n      restartPolicy: Never\n      containers<\/span>:\n        - name: task\n          image<\/span>: busybox:1.36<\/span>\n          command<\/span>:\n            - \/bin\/<\/span>sh\n            - -c\n            - |\n              echo \"This Job will be cleaned up after 60 seconds\"<\/span>\n              date\nEOF\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-ttl<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> jobs,pods\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
After about 60 seconds from completion, Kubernetes becomes eligible to delete the Job and its Pods.<\/p>\n\n\n\n
\n\n\n\n
13. Parallel Jobs<\/h1>\n\n\n\n
Some tasks can run in parallel.<\/p>\n\n\n\n
Example use cases:<\/p>\n\n\n
Process 100 files\nSend 10,000 emails\nRun 50 test cases\nProcess data partitions\nRun simulations\n<\/code><\/span><\/pre>\n\n\n
Kubernetes supports parallel Jobs using:<\/p>\n\n\n
completions<\/span>: 6\nparallelism<\/span>: 2\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Meaning:<\/p>\n\n\n
Need 6 successful Pods total.\nRun maximum 2 Pods at a time.\n<\/code><\/span><\/pre>\n\n\n
Create:<\/p>\n\n\n
cat > job-parallel.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: parallel-job\nspec:\n  completions: 6<\/span>\n  parallelism: 2<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: worker\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Worker started on $(hostname)\"<\/span>\n              sleep 10<\/span>\n              echo<\/span> \"Worker completed on $(hostname)\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-parallel<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> pods -l job-name=parallel-job -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check Job:<\/p>\n\n\n
kubectl get<\/span> job parallel-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Expected:<\/p>\n\n\n
COMPLETIONS\n6\/6\n<\/code><\/span><\/pre>\n\n\n
Kubernetes documentation states that .spec.parallelism<\/code> controls the requested number of Pods running at any instant, and .spec.completions<\/code> controls how many successful Pods are required for fixed-completion Jobs.<\/p>\n\n\n\n
\n\n\n\n
14. Indexed Jobs<\/h1>\n\n\n\n
An Indexed Job gives each Pod a unique completion index.<\/p>\n\n\n\n
This is useful when each Pod must process a specific shard or partition:<\/p>\n\n\n
Pod index 0 processes file group 0\nPod index 1 processes file group 1\nPod index 2 processes file group 2\n<\/code><\/span><\/pre>\n\n\n
Kubernetes supports completionMode: Indexed<\/code>. For Indexed Jobs, each Pod gets an index from 0<\/code> to .spec.completions - 1<\/code>, and the Job is complete when one Pod succeeds for every index. The index is available through mechanisms including annotation, label, hostname, and the JOB_COMPLETION_INDEX<\/code> environment variable.<\/p>\n\n\n\n
Create:<\/p>\n\n\n
cat > job-indexed.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: indexed-job\nspec:\n  completions: 5<\/span>\n  parallelism: 2<\/span>\n  completionMode: Indexed\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: indexed-worker\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"This is worker index: ${JOB_COMPLETION_INDEX}\"<\/span>\n              echo<\/span> \"Hostname: $(hostname)\"<\/span>\n              sleep 5<\/span>\n              echo<\/span> \"Completed index: ${JOB_COMPLETION_INDEX}\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-indexed<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check Pods:<\/p>\n\n\n
kubectl get<\/span> pods -l job-name=indexed-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check logs:<\/p>\n\n\n
kubectl logs -l job-name=indexed-job --prefix=true<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Expected style:<\/p>\n\n\n
[pod\/indexed-job-0-xxxxx\/indexed-worker]<\/span> This<\/span> is<\/span> worker<\/span> index<\/span>: 0\n[pod\/indexed-job-1-yyyyy\/indexed-worker]<\/span> This<\/span> is<\/span> worker<\/span> index<\/span>: 1\n[pod\/indexed-job-2-zzzzz\/indexed-worker]<\/span> This<\/span> is<\/span> worker<\/span> index<\/span>: 2\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
15. Per-Index Retry with backoffLimitPerIndex<\/code><\/h1>\n\n\n\n
For Indexed Jobs, Kubernetes supports backoffLimitPerIndex<\/code>. This lets each index have its own retry budget. The Job API says this field can only be used when completionMode: Indexed<\/code> and the Pod restart policy is Never<\/code>.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
cat > job-backoff-per-index.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: backoff-per-index-job\nspec:\n  completions: 4<\/span>\n  parallelism: 2<\/span>\n  completionMode: Indexed\n  backoffLimitPerIndex: 1<\/span>\n  maxFailedIndexes: 2<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: worker\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Running index: ${JOB_COMPLETION_INDEX}\"<\/span>\n              if<\/span> [ \"${JOB_COMPLETION_INDEX}\"<\/span> = \"0\"<\/span> ]; then\n                echo<\/span> \"Index 0 fails intentionally\"<\/span>\n                exit<\/span> 1<\/span>\n              fi\n              echo<\/span> \"Index ${JOB_COMPLETION_INDEX} completed\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-backoff-per-index<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> job backoff-per-index-job -o yaml\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Useful fields:<\/p>\n\n\n
kubectl get job backoff-per-index-job \\\n  -o jsonpath='{.status.completedIndexes}{\"\\n\"}{.status.failedIndexes}{\"\\n\"}'<\/span>\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
16. Suspend and Resume a Job<\/h1>\n\n\n\n
A Job can be created in suspended mode. When suspended, the Job controller does not create Pods. If a running Job is suspended, Kubernetes deletes active Pods associated with that Job. The API docs also state that suspending a Job resets the Job start time and effectively resets the activeDeadlineSeconds<\/code> timer.<\/p>\n\n\n\n
Create suspended Job:<\/p>\n\n\n
cat > job-suspend.yaml <<'EOF'<\/span>\napiVersion<\/span>: batch\/v1\nkind<\/span>: Job\nmetadata<\/span>:\n  name: suspended-job\nspec<\/span>:\n  suspend: true<\/span>\n  template<\/span>:\n    spec:\n      restartPolicy: Never\n      containers<\/span>:\n        - name: task\n          image<\/span>: busybox:1.36<\/span>\n          command<\/span>:\n            - \/bin\/<\/span>sh\n            - -c\n            - |\n              echo \"This starts only after resume\"<\/span>\n              date\nEOF\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-suspend<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> jobs,pods\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
No Pod should be created yet.<\/p>\n\n\n\n
Resume:<\/p>\n\n\n
kubectl patch job suspended-job -p '{\"spec\":{\"suspend\":false}}'<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> pods -l job-name=suspended-job -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
17. Create a Job Quickly Using Command Line<\/h1>\n\n\n\n
You can create a Job without writing YAML.<\/p>\n\n\n
kubectl create job quick-job \\\n  --image=busybox:1.36<\/span> \\\n  -- \/bin\/<\/span>sh -c 'echo \"Quick Job\"; date'<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> jobs\nkubectl logs job\/quick-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc create job quick-job \\\n  --image=busybox:1.36<\/span> \\\n  -- \/bin\/<\/span>sh -c 'echo \"Quick Job\"; date'<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
18. CronJob Schedule Syntax<\/h1>\n\n\n\n
CronJob uses standard cron-style syntax:<\/p>\n\n\n
# \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 minute        0 - 59<\/span>\n# \u2502 \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 hour        0 - 23<\/span>\n# \u2502 \u2502 \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 day       1 - 31<\/span>\n# \u2502 \u2502 \u2502 \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 month   1 - 12<\/span>\n# \u2502 \u2502 \u2502 \u2502 \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 week  0 - 6, Sunday to Saturday<\/span>\n# \u2502 \u2502 \u2502 \u2502 \u2502<\/span>\n# * * * * *<\/span>\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Examples:<\/p>\n\n\n\n
Schedule<\/th>Meaning<\/th><\/tr><\/thead>
* * * * *<\/code><\/td>Every minute<\/td><\/tr>
*\/5 * * * *<\/code><\/td>Every 5 minutes<\/td><\/tr>
0 * * * *<\/code><\/td>Every hour<\/td><\/tr>
0 2 * * *<\/code><\/td>Every day at 2 AM<\/td><\/tr>
0 3 * * 1<\/code><\/td>Every Monday at 3 AM<\/td><\/tr>
0 0 1 * *<\/code><\/td>First day of every month<\/td><\/tr>
@hourly<\/code><\/td>Once per hour<\/td><\/tr>
@daily<\/code><\/td>Once per day<\/td><\/tr>
@weekly<\/code><\/td>Once per week<\/td><\/tr>
@monthly<\/code><\/td>Once per month<\/td><\/tr>
@yearly<\/code><\/td>Once per year<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Kubernetes CronJob requires .spec.schedule<\/code> and supports standard cron syntax, step values such as *\/2<\/code>, and macros such as @hourly<\/code>, @daily<\/code>, @weekly<\/code>, and @monthly<\/code>.<\/p>\n\n\n\n
\n\n\n\n
19. Basic CronJob Example<\/h1>\n\n\n\n
Create a CronJob that runs every minute.<\/p>\n\n\n
cat > cronjob-hello.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: CronJob\nmetadata:\n  name: hello-cronjob\nspec:\n  schedule: \"* * * * *\"<\/span>\n  jobTemplate:\n    spec:\n      template:\n        spec:\n          restartPolicy: OnFailure\n          containers:\n            - name: hello\n              image: busybox:1.36<\/span>\n              command:\n                - \/bin\/sh\n                - -c\n                - |\n                  echo<\/span> \"Hello from Kubernetes CronJob\"<\/span>\n                  echo<\/span> \"Run time:\"<\/span>\n                  date\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-hello<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check CronJob:<\/p>\n\n\n
kubectl get<\/span> cronjobs\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Expected:<\/p>\n\n\n
NAME            SCHEDULE    SUSPEND   ACTIVE   LAST SCHEDULE   AGE\nhello-cronjob   * * * * *   False     0        <none<\/span>><\/span>          10s\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Wait one minute, then check Jobs:<\/p>\n\n\n
kubectl get<\/span> jobs\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Expected:<\/p>\n\n\n
NAME                       STATUS     COMPLETIONS   DURATION   AGE\nhello-cronjob-xxxxxxxx     Complete   1\/1           3s         30s\n<\/code><\/span><\/pre>\n\n\n
Check Pods:<\/p>\n\n\n
kubectl get<\/span> pods\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check logs:<\/p>\n\n\n
kubectl logs job\/$(kubectl get jobs --sort-by=.metadata.creationTimestamp -o jsonpath='{.items[-1:].metadata.name}'<\/span>)\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc logs job\/$(oc get jobs --sort-by=.metadata.creationTimestamp -o jsonpath='{.items[-1:].metadata.name}'<\/span>)\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
20. Important CronJob Fields<\/h1>\n\n\n\n
Field<\/th>Meaning<\/th><\/tr><\/thead>
schedule<\/code><\/td>Required cron schedule<\/td><\/tr>
jobTemplate<\/code><\/td>Required Job template<\/td><\/tr>
timeZone<\/code><\/td>Time zone for schedule<\/td><\/tr>
concurrencyPolicy<\/code><\/td>Allow, Forbid, or Replace overlapping Jobs<\/td><\/tr>
startingDeadlineSeconds<\/code><\/td>Deadline for starting missed Jobs<\/td><\/tr>
suspend<\/code><\/td>Pause future schedules<\/td><\/tr>
successfulJobsHistoryLimit<\/code><\/td>Number of successful Jobs to keep<\/td><\/tr>
failedJobsHistoryLimit<\/code><\/td>Number of failed Jobs to keep<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
The CronJob API documents concurrencyPolicy<\/code>, failedJobsHistoryLimit<\/code>, jobTemplate<\/code>, schedule<\/code>, startingDeadlineSeconds<\/code>, successfulJobsHistoryLimit<\/code>, suspend<\/code>, and timeZone<\/code>. The defaults include concurrencyPolicy: Allow<\/code>, successfulJobsHistoryLimit: 3<\/code>, failedJobsHistoryLimit: 1<\/code>, and suspend: false<\/code>.<\/p>\n\n\n\n
\n\n\n\n
21. CronJob with Time Zone<\/h1>\n\n\n\n
Kubernetes CronJobs support .spec.timeZone<\/code>, stable since Kubernetes v1.27. If no time zone is specified, the controller interprets schedules relative to the local time zone of kube-controller-manager<\/code>. Kubernetes recommends using the timeZone<\/code> field rather than putting TZ<\/code> or CRON_TZ<\/code> inside the schedule.<\/p>\n\n\n\n
Example: run at 9 AM Tokyo time.<\/p>\n\n\n
cat > cronjob-timezone.yaml <<'EOF'<\/span>\napiVersion<\/span>: batch\/v1\nkind<\/span>: CronJob\nmetadata<\/span>:\n  name: tokyo-cronjob\nspec<\/span>:\n  schedule: \"0 9 * * *\"<\/span>\n  timeZone<\/span>: \"Asia\/Tokyo\"<\/span>\n  jobTemplate<\/span>:\n    spec:\n      template:\n        spec:\n          restartPolicy: OnFailure\n          containers<\/span>:\n            - name: tokyo-task\n              image<\/span>: busybox:1.36<\/span>\n              command<\/span>:\n                - \/bin\/<\/span>sh\n                - -c\n                - |\n                  echo \"This runs daily at 9 AM Asia\/Tokyo\"<\/span>\n                  date\nEOF\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-timezone<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> cronjob tokyo-cronjob -o yaml\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
22. CronJob Concurrency Policy<\/h1>\n\n\n\n
Sometimes a scheduled task takes longer than the interval between runs.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
CronJob runs every minute.\nEach Job takes 5 minutes.\nShould Kubernetes start overlapping Jobs?\n<\/code><\/span><\/pre>\n\n\n
This is controlled by:<\/p>\n\n\n
concurrencyPolicy<\/span>: Allow\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Available values:<\/p>\n\n\n\n
Policy<\/th>Meaning<\/th><\/tr><\/thead>
Allow<\/code><\/td>Default. Allows overlapping Jobs<\/td><\/tr>
Forbid<\/code><\/td>Skips new run if previous run is still active<\/td><\/tr>
Replace<\/code><\/td>Cancels currently running Job and starts a new one<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Kubernetes documentation says Allow<\/code> permits concurrent Jobs, Forbid<\/code> skips a new run if the previous one is still running, and Replace<\/code> cancels the current run and replaces it with a new run. This policy applies only to Jobs created by the same CronJob.<\/p>\n\n\n\n
\n\n\n\n
22.1 Example: concurrencyPolicy: Forbid<\/code><\/h2>\n\n\n\n
This CronJob runs every minute, but each Job sleeps for 90 seconds. With Forbid<\/code>, Kubernetes will skip overlapping executions.<\/p>\n\n\n
cat > cronjob-forbid.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: CronJob\nmetadata:\n  name: forbid-cronjob\nspec:\n  schedule: \"* * * * *\"<\/span>\n  concurrencyPolicy: Forbid\n  jobTemplate:\n    spec:\n      template:\n        spec:\n          restartPolicy: Never\n          containers:\n            - name: slow-task\n              image: busybox:1.36<\/span>\n              command:\n                - \/bin\/sh\n                - -c\n                - |\n                  echo<\/span> \"Started at:\"<\/span>\n                  date\n                  sleep 90<\/span>\n                  echo<\/span> \"Finished at:\"<\/span>\n                  date\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-forbid<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> cronjob forbid-cronjob -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check Jobs:<\/p>\n\n\n
kubectl get<\/span> jobs --sort-by=.metadata.creationTimestamp\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
You should see fewer Jobs than one per minute because overlapping runs are forbidden.<\/p>\n\n\n\n
\n\n\n\n
22.2 Example: concurrencyPolicy: Replace<\/code><\/h2>\n\n\n
cat > cronjob-replace.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: CronJob\nmetadata:\n  name: replace-cronjob\nspec:\n  schedule: \"* * * * *\"<\/span>\n  concurrencyPolicy: Replace\n  jobTemplate:\n    spec:\n      template:\n        spec:\n          restartPolicy: Never\n          containers:\n            - name: replace-task\n              image: busybox:1.36<\/span>\n              command:\n                - \/bin\/sh\n                - -c\n                - |\n                  echo<\/span> \"Started at:\"<\/span>\n                  date\n                  sleep 120<\/span>\n                  echo<\/span> \"Finished at:\"<\/span>\n                  date\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-replace<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch Jobs:<\/p>\n\n\n
kubectl get<\/span> jobs,pods -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Every new scheduled run can replace the currently running one.<\/p>\n\n\n\n
\n\n\n\n
23. startingDeadlineSeconds<\/code><\/h1>\n\n\n\n
startingDeadlineSeconds<\/code> tells Kubernetes how late a missed CronJob run may start. If the delay is greater than the configured deadline, Kubernetes skips that occurrence. Kubernetes treats missed Jobs beyond the deadline as failed Jobs.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
startingDeadlineSeconds<\/span>: 300\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Meaning:<\/p>\n\n\n
If the Job cannot start within 300 seconds of scheduled time,\nskip that occurrence.\n<\/code><\/span><\/pre>\n\n\n
Example manifest:<\/p>\n\n\n
cat > cronjob-deadline.yaml <<'EOF'<\/span>\napiVersion<\/span>: batch\/v1\nkind<\/span>: CronJob\nmetadata<\/span>:\n  name: deadline-cronjob\nspec<\/span>:\n  schedule: \"*\/5 * * * *\"<\/span>\n  startingDeadlineSeconds<\/span>: 120<\/span>\n  concurrencyPolicy<\/span>: Forbid\n  jobTemplate<\/span>:\n    spec:\n      template:\n        spec:\n          restartPolicy: OnFailure\n          containers<\/span>:\n            - name: deadline-task\n              image<\/span>: busybox:1.36<\/span>\n              command<\/span>:\n                - \/bin\/<\/span>sh\n                - -c\n                - |\n                  echo \"CronJob with startingDeadlineSeconds\"<\/span>\n                  date\nEOF\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-deadline<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
24. Suspend and Resume CronJob<\/h1>\n\n\n\n
You can pause future CronJob executions using:<\/p>\n\n\n
suspend: true<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
This does not stop already started Jobs. Kubernetes documentation states that suspend<\/code> affects subsequent executions, not Jobs already started. It also warns that suspended executions count as missed Jobs, and when unsuspended without a starting deadline, missed Jobs may be scheduled immediately.<\/p>\n\n\n\n
Suspend:<\/p>\n\n\n
kubectl patch cronjob hello-cronjob -p '{\"spec\":{\"suspend\":true}}'<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> cronjob hello-cronjob\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Resume:<\/p>\n\n\n
kubectl patch cronjob hello-cronjob -p '{\"spec\":{\"suspend\":false}}'<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Better production pattern:<\/p>\n\n\n
startingDeadlineSeconds<\/span>: 300\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Use this with suspend\/resume to avoid accidental burst execution after long suspension.<\/p>\n\n\n\n
\n\n\n\n
25. CronJob History Limits<\/h1>\n\n\n\n
CronJobs can keep old successful and failed Jobs.<\/p>\n\n\n\n
Defaults:<\/p>\n\n\n
successfulJobsHistoryLimit<\/span>: 3\nfailedJobsHistoryLimit<\/span>: 1\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Set to lower values to avoid clutter.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
cat > cronjob-history.yaml <<'EOF'<\/span>\napiVersion<\/span>: batch\/v1\nkind<\/span>: CronJob\nmetadata<\/span>:\n  name: history-cronjob\nspec<\/span>:\n  schedule: \"* * * * *\"<\/span>\n  successfulJobsHistoryLimit<\/span>: 2<\/span>\n  failedJobsHistoryLimit<\/span>: 1<\/span>\n  jobTemplate<\/span>:\n    spec:\n      template:\n        spec:\n          restartPolicy: OnFailure\n          containers<\/span>:\n            - name: task\n              image<\/span>: busybox:1.36<\/span>\n              command<\/span>:\n                - \/bin\/<\/span>sh\n                - -c\n                - |\n                  echo \"CronJob history demo\"<\/span>\n                  date\nEOF\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-history<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Kubernetes documentation says successfulJobsHistoryLimit<\/code> controls how many successful Jobs are kept and defaults to 3<\/code>; failedJobsHistoryLimit<\/code> controls failed Jobs and defaults to 1<\/code>. Setting either to 0<\/code> keeps none of that type.<\/p>\n\n\n\n
\n\n\n\n
26. Manually Trigger a CronJob Immediately<\/h1>\n\n\n\n
A CronJob runs on schedule, but sometimes you want to run it now.<\/p>\n\n\n\n
Use:<\/p>\n\n\n
kubectl create job manual-run-1<\/span> --from<\/span>=cronjob\/hello-cronjob\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> jobs\nkubectl logs job\/manual-run-1\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc create job manual-run-1<\/span> --from<\/span>=cronjob\/hello-cronjob\noc logs job\/manual-run-1<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
This is very useful for testing backup CronJobs without waiting for the real schedule.<\/p>\n\n\n\n
\n\n\n\n
27. Real-World Use Case 1: Database Backup CronJob<\/h1>\n\n\n\n
This example uses busybox<\/code> for demonstration. In production, use a database-specific image such as PostgreSQL, MySQL, MongoDB, or your internal backup image.<\/p>\n\n\n\n
Create Secret:<\/p>\n\n\n
kubectl create secret generic db-backup-secret \\\n  --from<\/span>-literal=DB_HOST='mysql.default.svc.cluster.local'<\/span> \\\n  --from<\/span>-literal=DB_USER='backup_user'<\/span> \\\n  --from<\/span>-literal=DB_PASSWORD='backup_password'<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Create CronJob:<\/p>\n\n\n
cat > cronjob-db-backup.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: CronJob\nmetadata:\n  name: db-backup\nspec:\n  schedule: \"0 2 * * *\"<\/span>\n  timeZone: \"Etc\/UTC\"<\/span>\n  concurrencyPolicy: Forbid\n  startingDeadlineSeconds: 1800<\/span>\n  successfulJobsHistoryLimit: 3<\/span>\n  failedJobsHistoryLimit: 3<\/span>\n  jobTemplate:\n    spec:\n      backoffLimit: 2<\/span>\n      ttlSecondsAfterFinished: 86400<\/span>\n      template:\n        spec:\n          restartPolicy: Never\n          containers:\n            - name: backup\n              image: busybox:1.36<\/span>\n              env:\n                - name: DB_HOST\n                  valueFrom:\n                    secretKeyRef:\n                      name: db-backup-secret\n                      key: DB_HOST\n                - name: DB_USER\n                  valueFrom:\n                    secretKeyRef:\n                      name: db-backup-secret\n                      key: DB_USER\n                - name: DB_PASSWORD\n                  valueFrom:\n                    secretKeyRef:\n                      name: db-backup-secret\n                      key: DB_PASSWORD\n              command:\n                - \/bin\/sh\n                - -c\n                - |\n                  echo<\/span> \"Starting DB backup\"<\/span>\n                  echo<\/span> \"DB_HOST=${DB_HOST}\"<\/span>\n                  echo<\/span> \"Backup user=${DB_USER}\"<\/span>\n                  echo<\/span> \"Never print real password in production\"<\/span>\n                  date\n                  echo<\/span> \"Backup completed\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-db-backup<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Test now:<\/p>\n\n\n
kubectl create job db-backup-manual --from<\/span>=cronjob\/db-backup\nkubectl logs job\/db-backup-manual\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Production notes:<\/p>\n\n\n
Do<\/span> not print<\/span> secrets in logs.\nStore backups in object storage.\nUse<\/span> Workload<\/span> Identity<\/span> or<\/span> IAM<\/span> roles<\/span> where<\/span> possible<\/span>.\nUse<\/span> encryption<\/span> for<\/span> backups<\/span>.\nUse<\/span> concurrencyPolicy<\/span>: Forbid<\/span> for<\/span> backups<\/span>.\nUse<\/span> startingDeadlineSeconds<\/span> to<\/span> prevent<\/span> stale<\/span> backups<\/span>.\nUse<\/span> resource<\/span> requests<\/span> and<\/span> limits<\/span>.\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
28. Real-World Use Case 2: Report Generation CronJob<\/h1>\n\n\n
cat > cronjob-report.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: CronJob\nmetadata:\n  name: daily-report\nspec:\n  schedule: \"30 6 * * *\"<\/span>\n  timeZone: \"Asia\/Tokyo\"<\/span>\n  concurrencyPolicy: Forbid\n  successfulJobsHistoryLimit: 5<\/span>\n  failedJobsHistoryLimit: 2<\/span>\n  jobTemplate:\n    spec:\n      backoffLimit: 3<\/span>\n      ttlSecondsAfterFinished: 172800<\/span>\n      template:\n        spec:\n          restartPolicy: Never\n          containers:\n            - name: report\n              image: busybox:1.36<\/span>\n              command:\n                - \/bin\/sh\n                - -c\n                - |\n                  echo<\/span> \"Generating daily report\"<\/span>\n                  date\n                  sleep 10<\/span>\n                  echo<\/span> \"Report generated successfully\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-report<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> cronjob daily-report\nkubectl describe cronjob daily-report\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Manual run:<\/p>\n\n\n
kubectl create job daily-report-test --from<\/span>=cronjob\/daily-report\nkubectl logs job\/daily-report-test\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
29. Real-World Use Case 3: Cleanup Job<\/h1>\n\n\n\n
Use a Job for one-time cleanup.<\/p>\n\n\n
cat > job-cleanup.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: one-time-cleanup\nspec:\n  ttlSecondsAfterFinished: 300<\/span>\n  backoffLimit: 1<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: cleanup\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Running one-time cleanup\"<\/span>\n              date\n              echo<\/span> \"Cleanup completed\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-cleanup<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> job one-time-cleanup\nkubectl logs job\/one-time-cleanup\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
30. Real-World Use Case 4: Parallel File Processing<\/h1>\n\n\n
cat > job-file-processing.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: file-processing\nspec:\n  completions: 10<\/span>\n  parallelism: 3<\/span>\n  completionMode: Indexed\n  backoffLimitPerIndex: 2<\/span>\n  maxFailedIndexes: 3<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: processor\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Processing file partition: ${JOB_COMPLETION_INDEX}\"<\/span>\n              sleep 5<\/span>\n              echo<\/span> \"Completed partition: ${JOB_COMPLETION_INDEX}\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-file-processing<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> pods -l job-name=file-processing -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Logs:<\/p>\n\n\n
kubectl logs -l job-name=file-processing --prefix=true<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
31. Advanced: Pod Failure Policy<\/h1>\n\n\n\n
podFailurePolicy<\/code> gives more precise control over failures than only using backoffLimit<\/code>. Kubernetes documentation says Pod failure policy became stable in Kubernetes v1.31 and lets the Job handle failures based on container exit codes and Pod conditions.<\/p>\n\n\n\n
Example: fail the whole Job immediately if the container exits with code 42<\/code>.<\/p>\n\n\n
cat > job-pod-failure-policy.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: pod-failure-policy-job\nspec:\n  completions: 3<\/span>\n  parallelism: 1<\/span>\n  backoffLimit: 6<\/span>\n  podFailurePolicy:\n    rules:\n      - action: FailJob\n        onExitCodes:\n          containerName: main\n          operator: In\n          values: [42<\/span>]\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: main\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Simulating non-retriable application bug\"<\/span>\n              exit<\/span> 42<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-pod-failure-policy<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl describe job pod-failure-policy-job\n<\/code><\/span><\/pre>\n\n\n
Expected: the Job fails quickly instead of wasting retries.<\/p>\n\n\n\n
Important limitation: the Job API says podFailurePolicy<\/code> cannot be used together with restartPolicy: OnFailure<\/code>. Use restartPolicy: Never<\/code>.<\/p>\n\n\n\n
\n\n\n\n
32. Advanced: Success Policy for Indexed Jobs<\/h1>\n\n\n\n
For some batch workloads, not every Pod must succeed. For example:<\/p>\n\n\n
Run 10 simulations.\nIf any 3 important simulations succeed, mark the Job successful.\n<\/code><\/span><\/pre>\n\n\n
Kubernetes supports successPolicy<\/code> for Indexed Jobs. The documentation says success policy lets you declare a Job successful based on succeeded indexes or succeeded count, and after the policy is met, lingering Pods are terminated.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
cat > job-success-policy.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: success-policy-job\nspec:\n  parallelism: 5<\/span>\n  completions: 5<\/span>\n  completionMode: Indexed\n  successPolicy:\n    rules:\n      - succeededIndexes: 0<\/span>-4<\/span>\n        succeededCount: 3<\/span>\n  template:\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: worker\n          image: busybox:1.36<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Worker index: ${JOB_COMPLETION_INDEX}\"<\/span>\n              sleep 5<\/span>\n              echo<\/span> \"Success from index ${JOB_COMPLETION_INDEX}\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-success-policy<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch:<\/p>\n\n\n
kubectl get<\/span> job success-policy-job -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
33. Debugging Jobs and CronJobs<\/h1>\n\n\n\n
Check Jobs<\/h2>\n\n\n
kubectl get jobs\nkubectl describe job <job-name<\/span>><\/span>\nkubectl get job <job-name<\/span>><\/span> -o yaml\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Check Pods created by a Job<\/h2>\n\n\n
kubectl get pods -l job-name=<job-name<\/span>><\/span>\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
View logs<\/h2>\n\n\n
kubectl logs job\/<job-name<\/span>><\/span>\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
If multiple Pods exist:<\/p>\n\n\n
kubectl logs -l job-name=<job-name<\/span>><\/span> --prefix=true\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Check CronJobs<\/h2>\n\n\n
kubectl get cronjobs\nkubectl describe cronjob <cronjob-name<\/span>><\/span>\nkubectl get cronjob <cronjob-name<\/span>><\/span> -o yaml\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Check Jobs created by CronJob<\/h2>\n\n\n
kubectl get<\/span> jobs --sort-by=.metadata.creationTimestamp\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch everything<\/h2>\n\n\n
kubectl get<\/span> cronjobs,jobs,pods -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Check events<\/h2>\n\n\n
kubectl get<\/span> events --sort-by=.lastTimestamp\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc get cronjobs,jobs,pods\noc describe cronjob <cronjob-name<\/span>><\/span>\noc logs job\/<job-name<\/span>><\/span>\noc get events --sort-by=.lastTimestamp\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
34. Troubleshooting Common Problems<\/h1>\n\n\n\n
Problem 1: Job keeps failing<\/h2>\n\n\n\n
Check:<\/p>\n\n\n
kubectl describe job <job-name<\/span>><\/span>\nkubectl get pods -l job-name=<job-name<\/span>><\/span>\nkubectl logs -l job-name=<job-name<\/span>><\/span> --prefix=true\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Common causes:<\/p>\n\n\n
Application exits with<\/span> non-zero code\nWrong command or args\nMissing ConfigMap or Secret\nImagePullBackOff\nPermission issue\nNetwork dependency unavailable\nResource limit too low\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
Problem 2: CronJob is not creating Jobs<\/h2>\n\n\n\n
Check:<\/p>\n\n\n
kubectl describe cronjob <cronjob-name<\/span>><\/span>\nkubectl get events --sort-by=.lastTimestamp\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Possible reasons:<\/p>\n\n\n
CronJob is suspended\nSchedule is wrong\nTime zone is wrong\nstartingDeadlineSeconds skipped the run\nController manager issue\nRBAC\/admission issue\n<\/code><\/span><\/pre>\n\n\n
Check suspend:<\/p>\n\n\n
kubectl get cronjob <cronjob-name<\/span>><\/span> -o jsonpath='{.spec.suspend}{\"\\n\"}'\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Resume:<\/p>\n\n\n
kubectl patch cronjob <cronjob-name<\/span>><\/span> -p '{\"spec\":{\"suspend\":false}}'\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
Problem 3: Too many old Jobs<\/h2>\n\n\n\n
Use history limits:<\/p>\n\n\n
successfulJobsHistoryLimit<\/span>: 2\nfailedJobsHistoryLimit<\/span>: 1\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Or Job TTL:<\/p>\n\n\n
ttlSecondsAfterFinished<\/span>: 3600\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Kubernetes TTL cleanup applies to Jobs and can delete finished Jobs and their dependent objects after the TTL expires.<\/p>\n\n\n\n
\n\n\n\n
Problem 4: CronJob creates overlapping Jobs<\/h2>\n\n\n\n
Use:<\/p>\n\n\n
concurrencyPolicy<\/span>: Forbid\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
For backups, cleanup tasks, and reports, Forbid<\/code> is usually safer than Allow<\/code>.<\/p>\n\n\n\n
\n\n\n\n
Problem 5: CronJob runs in wrong time zone<\/h2>\n\n\n\n
Use:<\/p>\n\n\n
timeZone: \"Asia\/Tokyo\"<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Do not use:<\/p>\n\n\n
schedule: \"CRON_TZ=Asia\/Tokyo 0 9 * * *\"<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Kubernetes documentation says CRON_TZ<\/code> and TZ<\/code> inside .spec.schedule<\/code> are not officially supported and should not be used; use .spec.timeZone<\/code> instead.<\/p>\n\n\n\n
\n\n\n\n
35. Production Best Practices<\/h1>\n\n\n\n
35.1 Make Jobs idempotent<\/h2>\n\n\n\n
A Job might run more than once due to retries, node failures, or controller behavior. Kubernetes documentation explicitly notes that even with parallelism: 1<\/code>, completions: 1<\/code>, and restartPolicy: Never<\/code>, the same program may sometimes be started twice.<\/p>\n\n\n\n
Design your task so repeated execution does not corrupt data.<\/p>\n\n\n\n
Good patterns:<\/p>\n\n\n
Use<\/span> unique<\/span> transaction<\/span> IDs<\/span>\nUse<\/span> locks<\/span> carefully<\/span>\nCheck<\/span> whether<\/span> work<\/span> is<\/span> already<\/span> completed<\/span>\nWrite<\/span> output<\/span> atomically<\/span>\nUse<\/span> checkpoints<\/span>\nUse<\/span> idempotent<\/span> database<\/span> operations<\/span>\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
35.2 Avoid using latest<\/code> image tag in production<\/h2>\n\n\n\n
Bad:<\/p>\n\n\n
image<\/span>: myapp<\/span>:latest<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Better:<\/p>\n\n\n
image<\/span>: myapp<\/span>:1.2.3<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Best:<\/p>\n\n\n
image: myapp@sha256:<digest<\/span>><\/span>\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
35.3 Always set resource requests and limits<\/h2>\n\n\n\n
Example:<\/p>\n\n\n
resources:\n  requests:\n    cpu: \"100m\"<\/span>\n    memory<\/span>: \"128Mi\"<\/span>\n  limits<\/span>:\n    cpu: \"500m\"<\/span>\n    memory<\/span>: \"512Mi\"<\/span>\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
35.4 Use concurrencyPolicy: Forbid<\/code> for critical CronJobs<\/h2>\n\n\n\n
For backups, financial reports, cleanup tasks, or data processing, overlapping Jobs can create corruption.<\/p>\n\n\n\n
Recommended:<\/p>\n\n\n
concurrencyPolicy<\/span>: Forbid\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
35.5 Use startingDeadlineSeconds<\/code><\/h2>\n\n\n\n
This avoids very stale missed runs.<\/p>\n\n\n\n
Example:<\/p>\n\n\n
startingDeadlineSeconds<\/span>: 1800\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Meaning:<\/p>\n\n\n
Start only if the missed run is less than 30 minutes late.\n<\/code><\/span><\/pre>\n\n\n
\n\n\n\n
35.6 Use history limits and TTL<\/h2>\n\n\n\n
Recommended CronJob:<\/p>\n\n\n
successfulJobsHistoryLimit<\/span>: 3\nfailedJobsHistoryLimit<\/span>: 3\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Recommended Job template:<\/p>\n\n\n
ttlSecondsAfterFinished<\/span>: 86400\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
35.7 Do not print secrets in logs<\/h2>\n\n\n\n
Bad:<\/p>\n\n\n
echo<\/span> $DB_PASSWORD\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Good:<\/p>\n\n\n
echo<\/span> \"Using DB credentials from Secret\"<\/span>\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
35.8 Use ServiceAccount with least privilege<\/h2>\n\n\n\n
Example:<\/p>\n\n\n
serviceAccountName<\/span>: backup-runner\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
Give the ServiceAccount only the permissions required by the Job.<\/p>\n\n\n\n
\n\n\n\n
35.9 Use restartPolicy: Never<\/code> for clear failure tracking<\/h2>\n\n\n\n
For most batch jobs:<\/p>\n\n\n
restartPolicy<\/span>: Never\n<\/code><\/span>Code language:<\/span> HTTP<\/span> (<\/span>http<\/span>)<\/span><\/small><\/pre>\n\n\n
This makes retries easier to observe because each retry creates a separate Pod.<\/p>\n\n\n\n
\n\n\n\n
35.10 Add labels<\/h2>\n\n\n\n
Example:<\/p>\n\n\n
metadata:\n  labels:\n    app: reporting\n    workload-type: cronjob\n    owner: platform-team\n<\/code><\/span><\/pre>\n\n\n
This helps with:<\/p>\n\n\n
Monitoring\nCost tracking\nLog filtering\nTroubleshooting\nCleanup\n<\/code><\/span><\/pre>\n\n\n
\n\n\n\n
36. Complete Production-Style CronJob Template<\/h1>\n\n\n
cat > cronjob-production-template.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: CronJob\nmetadata:\n  name: production-cronjob\n  labels:\n    app: production-cronjob\n    owner: platform-team\nspec:\n  schedule: \"0 2 * * *\"<\/span>\n  timeZone: \"Etc\/UTC\"<\/span>\n  concurrencyPolicy: Forbid\n  startingDeadlineSeconds: 1800<\/span>\n  successfulJobsHistoryLimit: 3<\/span>\n  failedJobsHistoryLimit: 3<\/span>\n  suspend: false<\/span>\n  jobTemplate:\n    metadata:\n      labels:\n        app: production-cronjob\n        workload-type: batch\n    spec:\n      backoffLimit: 2<\/span>\n      activeDeadlineSeconds: 3600<\/span>\n      ttlSecondsAfterFinished: 86400<\/span>\n      template:\n        metadata:\n          labels:\n            app: production-cronjob\n        spec:\n          restartPolicy: Never\n          containers:\n            - name: task\n              image: busybox:1.36<\/span>\n              imagePullPolicy: IfNotPresent\n              resources:\n                requests:\n                  cpu: \"100m\"<\/span>\n                  memory: \"128Mi\"<\/span>\n                limits:\n                  cpu: \"500m\"<\/span>\n                  memory: \"512Mi\"<\/span>\n              command:\n                - \/bin\/sh\n                - -c\n                - |\n                  echo<\/span> \"Starting production batch task\"<\/span>\n                  date\n                  echo<\/span> \"Do real work here\"<\/span>\n                  echo<\/span> \"Completed production batch task\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> cronjob-production-template<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Test manually:<\/p>\n\n\n
kubectl create job production-cronjob-test --from<\/span>=cronjob\/production-cronjob\nkubectl logs job\/production-cronjob-test\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
37. Complete Production-Style Job Template<\/h1>\n\n\n
cat > job-production-template.yaml <<'EOF'<\/span>\napiVersion: batch\/v1\nkind: Job\nmetadata:\n  name: production-job\n  labels:\n    app: production-job\n    owner: platform-team\nspec:\n  backoffLimit: 2<\/span>\n  activeDeadlineSeconds: 1800<\/span>\n  ttlSecondsAfterFinished: 3600<\/span>\n  template:\n    metadata:\n      labels:\n        app: production-job\n    spec:\n      restartPolicy: Never\n      containers:\n        - name: task\n          image: busybox:1.36<\/span>\n          imagePullPolicy: IfNotPresent\n          resources:\n            requests:\n              cpu: \"100m\"<\/span>\n              memory: \"128Mi\"<\/span>\n            limits:\n              cpu: \"500m\"<\/span>\n              memory: \"512Mi\"<\/span>\n          command:\n            - \/bin\/sh\n            - -c\n            - |\n              echo<\/span> \"Starting one-time production job\"<\/span>\n              date\n              echo<\/span> \"Do real work here\"<\/span>\n              echo<\/span> \"Completed one-time production job\"<\/span>\nEOF\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply:<\/p>\n\n\n
kubectl<\/span> apply<\/span> -f<\/span> job-production-template<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
Check:<\/p>\n\n\n
kubectl get<\/span> job production-job\nkubectl logs job\/production-job\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
38. Job and CronJob Commands Cheat Sheet<\/h1>\n\n\n\n
Jobs<\/h2>\n\n\n
kubectl get jobs\nkubectl describe job <job-name<\/span>><\/span>\nkubectl get job <job-name<\/span>><\/span> -o yaml\nkubectl delete job <job-name<\/span>><\/span>\nkubectl logs job\/<job-name<\/span>><\/span>\nkubectl get pods -l job-name=<job-name<\/span>><\/span>\nkubectl logs -l job-name=<job-name<\/span>><\/span> --prefix=true\nkubectl create job <job-name<\/span>><\/span> --image=<image<\/span>><\/span> -- <command<\/span>><\/span>\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
CronJobs<\/h2>\n\n\n
kubectl get cronjobs\nkubectl describe cronjob <cronjob-name<\/span>><\/span>\nkubectl get cronjob <cronjob-name<\/span>><\/span> -o yaml\nkubectl delete cronjob <cronjob-name<\/span>><\/span>\nkubectl create job <manual-job-name<\/span>><\/span> --from=cronjob\/<cronjob-name<\/span>><\/span>\nkubectl patch cronjob <cronjob-name<\/span>><\/span> -p '{\"spec\":{\"suspend\":true}}'\nkubectl patch cronjob <cronjob-name<\/span>><\/span> -p '{\"spec\":{\"suspend\":false}}'\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Events<\/h2>\n\n\n
kubectl get<\/span> events --sort-by=.lastTimestamp\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Watch<\/h2>\n\n\n
kubectl get<\/span> cronjobs,jobs,pods -w\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
39. OpenShift oc<\/code> Cheat Sheet<\/h1>\n\n\n\n
Everything works similarly with oc<\/code>.<\/p>\n\n\n
oc get jobs\noc describe job <job-name<\/span>><\/span>\noc logs job\/<job-name<\/span>><\/span>\noc get cronjobs\noc describe cronjob <cronjob-name<\/span>><\/span>\noc create job manual-run --from=cronjob\/<cronjob-name<\/span>><\/span>\noc patch cronjob <cronjob-name<\/span>><\/span> -p '{\"spec\":{\"suspend\":true}}'\noc patch cronjob <cronjob-name<\/span>><\/span> -p '{\"spec\":{\"suspend\":false}}'\noc get events --sort-by=.lastTimestamp\n<\/code><\/span>Code language:<\/span> HTML, XML<\/span> (<\/span>xml<\/span>)<\/span><\/small><\/pre>\n\n\n
Create project:<\/p>\n\n\n
oc new<\/span>-project jobs-demo\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Apply YAML:<\/p>\n\n\n
oc<\/span> apply<\/span> -f<\/span> job-hello<\/span>.yaml<\/span>\noc<\/span> apply<\/span> -f<\/span> cronjob-hello<\/span>.yaml<\/span>\n<\/code><\/span>Code language:<\/span> CSS<\/span> (<\/span>css<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
40. Cleanup All Tutorial Resources<\/h1>\n\n\n
kubectl delete<\/span> namespace jobs-demo\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
For OpenShift:<\/p>\n\n\n
oc delete<\/span> project jobs-demo\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
Or delete individual resources:<\/p>\n\n\n
kubectl delete<\/span> job --all\nkubectl delete<\/span> cronjob --all\nkubectl delete<\/span> secret db-backup-secret\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
\n\n\n\n
41. Final Summary<\/h1>\n\n\n\n
Kubernetes Jobs and CronJobs are essential for running batch workloads.<\/p>\n\n\n\n
Use a Job<\/strong> when:<\/p>\n\n\n
You need to run a task once.\nYou need a migration.\nYou need a one-time cleanup.\nYou need batch processing.\nYou need parallel workers.\n<\/code><\/span><\/pre>\n\n\n
Use a CronJob<\/strong> when:<\/p>\n\n\n
You need scheduled execution.\nYou need periodic backups.\nYou need daily reports.\nYou need weekly cleanup.\nYou need recurring sync tasks.\n<\/code><\/span><\/pre>\n\n\n
The simplest Job:<\/p>\n\n\n
apiVersion: batch\/v1\nkind<\/span>: Job\nmetadata<\/span>:\n  name: hello-job\nspec<\/span>:\n  template:\n    spec:\n      restartPolicy: Never\n      containers<\/span>:\n        - name: hello\n          image<\/span>: busybox:1.36<\/span>\n          command<\/span>: [\"\/bin\/sh\"<\/span>, \"-c\"<\/span>, \"echo Hello from Job\"<\/span>]\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
The simplest CronJob:<\/p>\n\n\n
apiVersion: batch\/v1\nkind<\/span>: CronJob\nmetadata<\/span>:\n  name: hello-cronjob\nspec<\/span>:\n  schedule: \"* * * * *\"<\/span>\n  jobTemplate<\/span>:\n    spec:\n      template:\n        spec:\n          restartPolicy: OnFailure\n          containers<\/span>:\n            - name: hello\n              image<\/span>: busybox:1.36<\/span>\n              command<\/span>: [\"\/bin\/sh\"<\/span>, \"-c\"<\/span>, \"date; echo Hello from CronJob\"<\/span>]\n<\/code><\/span>Code language:<\/span> JavaScript<\/span> (<\/span>javascript<\/span>)<\/span><\/small><\/pre>\n\n\n
The most important production lesson:<\/p>\n\n\n
Jobs and CronJobs must be idempotent, observable, resource-limited, retry-controlled, and cleaned up automatically.\n<\/code><\/span><\/pre>\n\n\n
The golden rule:<\/p>\n\n\n
Deployment runs forever.\nJob runs once.\nCronJob runs Jobs on a schedule.\n<\/code><\/span><\/pre>","protected":false},"excerpt":{"rendered":"
Kubernetes Jobs and CronJobs Explained with Examples: Complete Hands-on Guide for DevOps Engineer Keywords 1. Introduction In Kubernetes, not every workload is a long-running application like a… <\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_joinchat":[],"footnotes":""},"categories":[4859,25688],"tags":[],"class_list":["post-20595","post","type-post","status-publish","format-standard","hentry","category-kubernetes","category-kubernetes-api-resources"],"_links":{"self":[{"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/posts\/20595","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/comments?post=20595"}],"version-history":[{"count":6,"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/posts\/20595\/revisions"}],"predecessor-version":[{"id":77371,"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/posts\/20595\/revisions\/77371"}],"wp:attachment":[{"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/media?parent=20595"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/categories?post=20595"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.devopsschool.com\/blog\/wp-json\/wp\/v2\/tags?post=20595"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}