Hands-On Microservices with Kubernetes

Kubernetes Jobs and CronJobs

Kubernetes deployments and services are all about creating a set of long-running pods that are supposed to run indefinitely. A Kubernetes Job is all about running one or more pods until one of them completes successfully. When you create a Job, it looks very much like a deployment, except that the restart policy should be Never.

Here is a Kubernetes Job that prints Yeah, it works in a Job!!! from Python:

apiVersion: batch/v1kind: Jobmetadata:  name: yeah-it-worksspec:  template:    spec:      containers:      - name: yeah-it-works        image: python:3.6-alpine        command: ["python",  "-c", "print('Yeah, it works in a Job!!!')"]      restartPolicy: Never

I can now run this Job, watch it complete, and check the logs, as follows:

$ kubectl create ...