Job (K8s)
A Kubernetes object for a task that should run to completion and then finish — a batch process, a migration, a one-off computation. Unlike a Deployment (meant to run forever), a Job counts a pod as done when it has succeeded.
Mechanic: a Job creates one or more pods and watches them until the specified number of successful completions (completions) is reached. Features: parallelism (run several pods at once to chew through the work faster), backoffLimit (how many retries on failure before the Job is considered failed), and various completion modes (one task, or an indexed work set). Use case: a database migration at deploy, batch ETL, generating a report, processing a queue of tasks. Distinct from a Deployment (an always-on service that restarts on exit) — a Job is meant to end. A CronJob is a Job run on a schedule. Pitfall: don't forget to clean up old completed Jobs (ttlSecondsAfterFinished) or they pile up. Fundamental for batch workloads in Kubernetes. Related to CronJob (K8s) and Job.