CronJobs

A Kubernetes CronJob creates jobs on a schedule. Use cronjobs on RemoteGPU for recurring batch work.

Typical examples include nightly data refreshes, periodic evaluation jobs, cleanup tasks, health reports, and scheduled inference preparation work.

Choose a workflow

Use the console when you want to create a scheduled workload quickly or review cronjob, job, pod, log, and event status in a namespace.

Use kubectl when your team manages scheduled work from manifests or CI. Apply full CronJob manifests through the namespace kubeconfig and Kubernetes exec API key auth.

When to use a cronjob

A cronjob is useful for:

• recurring tasks that follow a cron-style schedule
• schedules and job templates managed in version-controlled manifests
• recurring namespace maintenance such as cleanup, sync, or report generation
• workloads where each scheduled run should create a standard Kubernetes job

For one-time execution, use a Job instead.

Supported cronjob profile

Area	Supported behavior
Resource type	batch/v1 CronJob
Access path	Console creation, plus kubectl for manifest create and update
Schedule format	Standard Kubernetes cron schedule syntax
Job template policy	The embedded job template uses the same plan and workload policy rules as other workloads
Namespace model	The cronjob and the jobs it creates stay inside the same namespace boundary

Create a cronjob

Use Kubernetes > CronJobs when you want to launch a straightforward recurring batch workload without writing a manifest first.

The console create flow supports:

• namespace selection
• cronjob name
• plan selection
• container image
• schedule
• optional shell command

Use kubectl when you need manifest fields that are not exposed in the console form, such as concurrency policy, history retention, suspend, starting deadline, labels, annotations, environment variables, volumes, or advanced job template settings. The same namespace workload policy applies either way.

Operating cronjobs on RemoteGPU

• When managed with kubectl, the job template inside the cronjob should include the RemoteGPU runtime SKU label.
• Set concurrencyPolicy intentionally so overlapping runs do not surprise you.
• Set success and failure history limits so completed jobs do not accumulate indefinitely.
• CronJobs are batch resources. They do not replace the deployment-service-ingress model used for public HTTP apps.

What you can do from the console

The console is still useful whether the cronjob was created in the console or with kubectl:

• create the namespace
• create a cronjob from the console form
• download namespace kubeconfig
• review current cronjobs in Kubernetes > CronJobs
• inspect pods, events, and logs created by scheduled runs
• inspect related services or ingresses in the same namespace

Troubleshooting

Symptom	What to check
CronJob does not create jobs	Confirm the schedule is valid Kubernetes cron syntax and the cronjob is not suspended
Scheduled jobs fail	Inspect the created job, pod events, and container logs; the job template follows the same rules as a normal Job
Runs overlap unexpectedly	Set concurrencyPolicy intentionally, such as Forbid when a new run should wait for the previous one to finish
Completed jobs accumulate	Set successfulJobsHistoryLimit and failedJobsHistoryLimit in the cronjob manifest
Console-created cronjob needs advanced fields	Use kubectl for fields such as concurrency policy, suspend, starting deadline, history limits, labels, annotations, environment variables, and volumes

Read next

• Read Jobs for one-off batch work.
• Read Deployments for long-running services.
• Read Kubernetes overview for the full access model.