Managing Jobs

1. Creating Job

Field	Description	Example
apiVersion	API version to use	`batch/v1`
kind	Resource type	`Job`
metadata.name	Name of the Job	`my-job`
spec.template.spec.containers	List of containers to run	`[{ name: "worker", image: "myapp:latest" }]`
spec.completions	Number of successful completions required	`1`
spec.parallelism	Number of concurrent pods to run	`1`

Example: Creating a Job

apiVersion: batch/v1
kind: Job
metadata:
  name: my-job
spec:
  template:
    spec:
      containers:
      - name: worker
        image: myapp:latest
      restartPolicy: Never
  completions: 1
  parallelism: 1

2. Creating Job from CronJob

Field	Description	Example
spec.schedule	Cron expression defining when to run the job	`0 0 * * *`
spec.jobTemplate	Template for the Job to be created	`{ "spec": { "template": { "spec": { "containers": [ { "name": "my-container", "image": "my-image:latest" } ], "restartPolicy": "OnFailure" } } } }`
spec.startingDeadlineSeconds	Deadline in seconds for starting the job if it misses its scheduled time	`300`
spec.concurrencyPolicy	Policy for handling concurrent executions (Allow, Forbid, Replace)	`Forbid`
spec.suspend	Boolean to suspend the CronJob (true = suspended)	`false`

Example: Creating a CronJob that runs a Job daily

apiVersion: batch/v1
kind: CronJob
metadata:
  name: daily-job
spec:
  schedule: "0 0 * * *"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
          - name: my-container
            image: my-image:latest
          restartPolicy: OnFailure

Note: The CronJob creates a new Job instance based on the jobTemplate at each scheduled time.

3. Listing Jobs

Command	Description	Flags
kubectl get jobs	List all jobs in the current namespace	`-A` to list across all namespaces
kubectl get jobs -o wide	List jobs with additional details like node name and age	`-o wide` for extended output
kubectl get jobs --sort-by=.metadata.creationTimestamp	List jobs sorted by creation timestamp	`--sort-by` flag with field path

Example: Listing Jobs in Current Namespace

kubectl get jobs
NAME       COMPLETIONS   AGE
job1       1/1           5m
job2       0/1           2m

Example: Listing Jobs Across All Namespaces

kubectl get jobs -A
NAMESPACE   NAME       COMPLETIONS   AGE
default     job1       1/1           5m
kube-system job2       0/1           2m

Note: The COMPLETIONS column shows the number of successful completions and total desired completions (e.g., 1/1 means job is complete).

4. Describing Job Details

Property	Description	Example
metadata.name	Name of the Job	`batch-job`
metadata.namespace	Namespace where the Job is defined	`default`
spec.template.spec.containers	List of containers in the Pod template	`[ { "name": "my-container", "image": "alpine:latest" } ]`
spec.completions	Number of completed pods required to consider the Job successful	`3`
spec.parallelism	Number of pods that can run in parallel	`2`
spec.backoffLimit	Maximum number of retries before marking the Job as failed	`4`
status.conditions	Current conditions of the Job	`[ { "type": "Complete", "status": "True" } ]`
status.active	Number of active pods	`1`
status.succeeded	Number of succeeded pods	`2`
status.failed	Number of failed pods	`0`

Example: Describe Job Details

apiVersion: batch/v1
kind: Job
metadata:
  name: batch-job
  namespace: default
spec:
  template:
    spec:
      containers:
      - name: my-container
        image: alpine:latest
        command: ["/bin/sh", "-c", "echo 'Hello World'"]
      restartPolicy: Never
  completions: 3
  parallelism: 2
  backoffLimit: 4
status:
  active: 1
  succeeded: 2
  failed: 0

5. Setting Completions Count

Field	Description	Values	Notes
spec.completions	Number of successful finished pods required to consider the Job as completed	`integer`	Default is 1 if not specified
spec.parallelism	Number of pods that are allowed to run in parallel	`integer`	Defaults to 1 if not specified
spec.activeDeadlineSeconds	Maximum duration in seconds for the Job to complete	`integer`	If exceeded, the Job is terminated

Example: Setting Completions Count

apiVersion: batch/v1
kind: Job
metadata:
  name: pi
spec:
  completions: 3
  parallelism: 2
  template:
    spec:
      containers:
      - name: pi
        image: perl
        command: ["perl",  "-Mbignum=bpi", "-wle", "print bpi(2000)"]
      restartPolicy: Never

6. Setting Parallelism Level

Field	Description	Values	Notes
parallelism	Number of pods running simultaneously	`integer`	Default is 1. Set to 0 to suspend job
completions	Total number of pods to run	`integer`	Must be >= parallelism
backoffLimit	Maximum number of retries before marking job as failed	`integer`	Default is 6

Example: Setting parallelism to 3

apiVersion: batch/v1
kind: Job
metadata:
  name: example-job
spec:
  parallelism: 3
  completions: 6
  template:
    spec:
      containers:
      - name: worker
        image: busybox
        command: ["echo", "Hello from job"]
      restartPolicy: Never

7. Configuring Backoff Limit

Property	Description	Default Value	Notes
backoffLimit	Number of retries before marking the job as failed	`6`	Set to `null` for infinite retries

Example: Setting Backoff Limit

apiVersion: batch/v1
kind: Job
metadata:
  name: example-job
spec:
  backoffLimit: 5
  template:
    spec:
      containers:
      - name: example-container
        image: busybox
        command: ["sh", "-c", "exit 1"]
      restartPolicy: Never

Note: Setting backoffLimit to null allows infinite retries, which can lead to resource exhaustion if not managed carefully.

8. Setting Active Deadline

Property	Value	Description
activeDeadlineSeconds	`300`	Maximum allowed duration in seconds for a Job to run before it is terminated

Example: Setting Active Deadline on a Job

apiVersion: batch/v1
kind: Job
metadata:
  name: example-job
spec:
  activeDeadlineSeconds: 300
  template:
    spec:
      containers:
      - name: example-container
        image: busybox
        command: ["sh", "-c", "echo Hello World"]
      restartPolicy: Never

9. Using Job Selectors

Selector Type	Description	Example
labelSelector	Matches jobs based on labels	`metadata.labels.environment: production`
fieldSelector	Matches jobs based on fields in the object's metadata	`status.phase: Failed`

Example: Label Selector

apiVersion: batch/v1
kind: Job
metadata:
  name: example-job
  labels:
    environment: production
    team: backend
spec:
  template:
    spec:
      containers:
      - name: job-container
        image: alpine:latest
        command: ["echo", "Hello World"]
      restartPolicy: Never

Example: Field Selector

apiVersion: batch/v1
kind: Job
metadata:
  name: example-job
spec:
  template:
    spec:
      containers:
      - name: job-container
        image: alpine:latest
        command: ["echo", "Hello World"]
      restartPolicy: Never

Note: Label selectors are more commonly used than field selectors in Kubernetes.

10. Viewing Job Logs

Command	Description	Notes
kubectl logs job/<job-name>	View logs from a completed job	Useful for debugging job failures
kubectl logs -l app=<app-name>	View logs for all pods matching a label	Useful when job has multiple pods
kubectl logs --previous job/<job-name>	View logs from previous job instance	Useful for troubleshooting failed jobs

Example: View job logs

kubectl logs job/my-job

Example: View logs from previous job run

kubectl logs --previous job/my-job

Note: Logs are only available while the pod is running or has not been garbage collected.

11. Deleting Jobs

Field	Description	Values
metadata.name	Name of the Job	`my-job`
metadata.namespace	Namespace where the Job is located	`default`
spec.completions	Number of successful completions required to consider the Job as complete	`1`
spec.parallelism	Number of pods that are allowed to run in parallel	`1`
spec.backoffLimit	Number of retries before marking the Job as failed	`4`

Example: Delete a Job by Name

kubectl delete job my-job

Example: Delete a Job in a Specific Namespace

kubectl delete job my-job -n production

Example: Delete All Jobs in a Namespace

kubectl delete jobs --all -n production

Note: Deleting a Job does not delete the Pods it created. You can use kubectl delete pods --selector=job-name=my-job to clean up the associated Pods.

Warning: Deleting a Job that is currently running will stop the execution of its Pods. Ensure that you have a backup or are okay with losing progress.

12. Cleaning Up Finished Jobs

Field	Description	Default Value
ttlSecondsAfterFinished	Number of seconds to retain job after completion	`null`

Example: Setting TTL for Job Cleanup

apiVersion: batch/v1
kind: Job
metadata:
  name: example-job
spec:
  ttlSecondsAfterFinished: 300
  template:
    spec:
      containers:
      - name: example-container
        image: busybox
        command: ["echo", "Hello World"]
      restartPolicy: Never

Note: Setting ttlSecondsAfterFinished allows automatic cleanup of completed jobs, preventing accumulation of finished resources in the cluster.

Warning: Once a job is cleaned up, its logs and status information are permanently lost. Ensure logs are collected or archived before cleanup.