Introduction

A CronJob is a declarative interface to describe time based Jobs.

API group Resource Kube Skeleton
batch/v1beta1 CronJob skel
batch/v2alpha1 CronJob skel

Here's an example Kubernetes CronJob:

apiVersion: batch/v1beta1
kind: CronJob
metadata:
  name: hello
spec:
  schedule: "*/1 * * * *"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
          - name: hello
            image: busybox
            args:
            - /bin/sh
            - -c
            - date; echo Hello from the Kubernetes cluster
          restartPolicy: OnFailure

The following sections contain detailed information about each field in Short syntax, including how the field translates to and from Kubernetes syntax.

API Overview

Field Type K8s counterpart(s) Description
version string apiVersion The version of the resource object
cluster string metadata.clusterName The name of the cluster on which this CronJob is running
name string metadata.name The name of the CronJob
namespace string metadata.namespace The K8s namespace this CronJob will be a member of
labels string metadata.labels Metadata about the CronJob, including identifying information
annotations string metadata.annotations Non-identifying information about the CronJob
schedule string spec.schedule The schedule in cron format (https://en.wikipedia.org/wiki/Cron)
suspend bool spec.suspend If set, then subsequent executions are suspended
start_deadline int64 spec.startingDeadlineSeconds Optional deadline for starting the cron-job if it misses the schedule. Missed cron-jobs are considered failed.
concurrency string spec.concurrencyPolicy Specifies how to treat concurrent executions of a cron-job. See Concurrency Policy
max_success_history int32 spec.successfulJobsHistoryLimit The number of successful finished cron-jobs to retain
max_failure_history int32 spec.failedJobsHistoryLimit The number of failed finished cron-jobs to retain
parallelism int32 spec.parallelism Maximum number of pods of this cron-job that can run in parallel
completions int32 spec.completions Minimum number of successfully completed pods for the cron-job to be considered successful
max_retries int32 spec.backOffLimit Maximum number of retries before considering this cron-job failed
active_deadline int32 spec.activeDeadlineSeconds Maximum time for a cron-job to be active before it is terminated by the system
select_manually bool spec.manualSelector Controls generation of pod labels and pod selectors. Defaults to false. If set, user is responsible for choosing pods correctly
selector map[string]string or string selector An expression (string) or a set of key, value pairs (map) that is used to select a set of pods to manage using the cron-job controller. See Selector Overview
pod_meta TemplateMetadata template Metadata of the Pod that is selected by this CronJob. See Template Metadata
volumes Volume spec.volumes Denotes the volumes that are a part of the Pod. See Volume Overview
affinity []Affinity spec.affinity and spec.NodeSelector The Pod's scheduling rules, expressed as (anti-)affinities for nodes or other Pods. See Affinity Overview
node string spec.nodeName Request that the Pod be scheduled on a specific node.
containers Container spec.containers and status Containers that run as a part of the Pod. See Container Overview
init_containers Container spec.initContainers and status Containers that run as a part of the initialization process of the Pod. See Container Overview
dns_policy DNSPolicy spec.dnsPolicy The DNS Policy of the Pod. See DNS Policy Overview
host_aliases []string spec.aliases Set of additional records to be placed in /etc/hosts file inside the Pod. See Host Aliases Overview
host_mode []string spec.hostPID, spec.hostNetwork and spec.hostIPC The Pod's access to host resources. See Host Mode Conversion
hostname string spec.hostname and spec.subDomain The fully qualified domain name of the pod
registry_secrets []string spec.ImagePullSecrets A list of k8s secret resource names that contain credentials to required to access private registries.
restart_policy RestartPolicy spec.restartPolicy Behavior of a Pod when it dies. Can be "always", "on-failure" or "never"
scheduler_name string spec.schedulerName The value from spec.schedulerName is stored here
account string spec.serviceAccountName and automountService AccountToken The Pod's access to the K8s API. See Account Conversion
tolerations []Toleration spec.tolerations Set of host taints this Pod tolerates. See Toleration Conversion
termination_ grace_period int64 spec.termination GracePeriodSeconds Number of seconds to wait before forcefully killing the Pod.
active_deadline int64 spec. activeDeadlineSeconds Number of seconds the Pod is allowed to be active
priority Priority spec.priorityClassName and spec.priority Specifies the Pod's Priority. See Priority
condition []Pod Condition status.conditions The list of current and previous conditions of the Pod. See Pod Condition
node_ip string status.hostIP The IP address of the Pod's host
ip string status.podIP The IP address of the Pod
start_time time status.startTime When the Pod started running
msg string status.message A human readable message explaining Pod's current condition
phase string status.phase The current phase of the Pod
reason string status.reason Reason indicating the cause for the current state of the Pod
qos string status.qosClass The QOS class assigned to the Pod based on resource requirements
fs_gid int64 spec.securityContext. fsGroup Special supplemental group that applies to all the Containers in the Pod
gids []int64 spec.securityContext. supplementalGroups A list of groups applied to the first process in each of the Containers in the Pod

Concurrency Policy

Concurrency Policy Description
allow Allows cron jobs to run concurrently
forbid forbids cron jobs from running concurrently, skipping next execution if the previous hasn't finished
replace replace cancels current job, to run the next one

Selector Overview

Selector can be a map value or a string value. If it is a string value, then it can be an expression of type

Valid Operators are

Operator Syntax Description
Eq = Key should be equal to value
Exists N/A Key should exist
NotExists N/A Key should not exist
In = Key should be one of the comma separated values
NotIn != Key should not be one of the comma separated values

Here are valid examples of all the expression operators

selector: key=value # key should be equal to value
selector: key # key should exist
selector: !key # key should not exist
selector: key=value1,value2 # key's value can be any of value1 or value2
selector: key!=value1,value2 # key's value cannot be any of value1 or value2
selector: key&key!=value # composite expression

Note that multiple expressions can be combined using the & symbol

If the selector is a map, then the values in the map are expected to match directly with the labels of a pod.

Template Metadata

Field Type K8s counterpart(s) Description
cluster string metadata.clusterName The name of the cluster on which this Pod is running
name string metadata.name The name of the Pod
namespace string metadata.namespace The K8s namespace this Pod will be a member of
labels string metadata.labels Metadata that could be identifying information about the Pod
annotations string metadata.annotations Non identifying information about the Pod

Examples

job:
  containers:
  - expose:
    - 80
    image: nginx:1.7.9
    name: nginx
  name: nginx-job
  parallelism: 1
  completions: 1
  schedule: '*/1 * * * *'
  selector:
    app: nginx
  version: batch/v1beta1
cron_job:
  containers:
  - expose:
    - 80
    image: nginx:1.7.9
    name: nginx
  name: nginx-job
  parallelism: 2
  completions: 4
  schedule: '*/1 * * * *'
  selector:
    app: nginx
  version: batch/v1beta1
cron_job:
  containers:
  - expose:
    - 80
    image: nginx:1.7.9
    name: nginx
  name: nginx-job
  parallelism: 2
  completions: 1
  schedule: '*/1 * * * *'
  selector:
    app: nginx
  version: batch/v1beta1

Skeleton

Short Type Skeleton
CronJob skel

Here's a starter skeleton of a Short CronJob.

cron_job:
  containers:
  - args:
    - /bin/sh
    - -c
    - date; echo Hello from the Kubernetes cluster
    image: busybox
    name: hello
  name: hello
  restart_policy: on-failure
  schedule: '*/1 * * * *'
  version: batch/v1beta1