README.md

Cronjobber

Cronjobber is the cronjob controller from Kubernetes patched with time zone support.

⚠️ Archived

Since Kubernetes 1.27, timezone support for CronJobs has been available as a stable feature, making this project obsolete. As a result, the project has been archived.

Thanks for all the fish, and happy scheduling!

Installation

# Install CustomResourceDefinition
$ kubectl apply -f https://raw.githubusercontent.com/hiddeco/cronjobber/master/deploy/crd.yaml
# Setup service account and RBAC
$ kubectl apply -f https://raw.githubusercontent.com/hiddeco/cronjobber/master/deploy/rbac.yaml
# Deploy Cronjobber (using the timezone db from the node)
$ kubectl apply -f https://raw.githubusercontent.com/hiddeco/cronjobber/master/deploy/deploy.yaml

Keeping your timezone database up-to-date

⚠️ Note: the approach below does not work at present due to an issue with Go which causes it to silently fail on slim timezone database formats (the default since tzdata>=2020-b) (go#42138).

It is therefore advised to use the deploy.yaml with the embedded timezone database until this issue has been resolved.

Cronjobber embeds a default timezone database in its binary, this database is however not updated regularly. To help your overcome this issue there is an cronjobber-updatetz image available that can be used as a sidecar.

# Deploy Cronjobber (using the updatetz sidecar)
$ kubectl apply -f https://raw.githubusercontent.com/hiddeco/cronjobber/master/deploy/deploy-updatetz.yaml

You may want to tweak the following environment variables to control how often it looks for updates and where it puts the timezone database.

• TZPV (default: /tmp/zoneinfo) is where the timezone database is extracted
• REFRESH_INTERVAL (default: 7d) is how often it looks for updates

Usage

Instead of creating a CronJob like you normally would, you create a TZCronJob, which works exactly the same but supports an additional field: .spec.timezone. Set this to the time zone you wish to schedule your jobs in and Cronjobber will take care of the rest.

apiVersion: cronjobber.hidde.co/v1alpha1
kind: TZCronJob
metadata:
  name: hello
spec:
  schedule: "*/1 * * * *"
  timezone: "Europe/Amsterdam"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
            - name: hello
              image: busybox
              args:
                - /bin/sh
                - -c
                - date; echo "Hello, World!"
          restartPolicy: OnFailure

Reasoning

There has been a long outstanding (and now closed) issue to add time zone support to the CronJob kind in Kubernetes, including a fully working PR which actually made it possible. SIG Apps and in SIG Architecture decided however against adding it because of the downside of having to manage and distribute time zone databases.

People are now encouraged to innovate and solve these kinds of problems in the ecosystem rather than core.

Instead of putting this in Kubernetes the ask is to:

1. Develop this in the ecosystem (e.g., a controller) that others can use. Distribute it, solve the problems there, and see what update looks like
2. If the solution is widely adopted and can be used by everyone (including small scale, multi-cluster, etc) then it could be considered for core Kubernetes

-- mattfarina (Matt Farina) on Jan 26, 2018

Cronjobber is the most simple answer to this: it is the original PR on top of a more recent version of the cronjob controller, with some glue added to make it an independent controller.

Credits

This application is derived from open source components. You can find the original source code of these components below.

• Kubernetes CronJob controller
• kubernetes/kubernetes#47266 by Adam Sunderland