1 Disclaimer
2 Version Information
3 Overview
- 3.1 Key Features
4 Requirements
5 Supported Actions
- 5.1 Action Output
6 Configuration Examples
7 Input Fields
8 Output Fields
9 Environment Variables
10 Cancelation and Rerun
11 Dynamic Commands
12 Exit Codes
13 STDOUT and STDERR
14 How To
- 14.1 Import Universal Template
- 14.2 Configure Universal Task
15 Integration Modifications
16 Document References
17 Changelog
- 17.1 ue-gke-jobs-2.1.0 (2026-01-22)
  - 17.1.1 Enhancements
  - 17.1.2 Fixes
- 17.2 ue-gke-jobs-2.0.0 (2025-06-26)
  - 17.2.1 Enhancements
  - 17.2.2 Breaking Changes
- 17.3 ue-gke-jobs-1.0.0 (2025-03-17)

Disclaimer

Your use of this download is governed by Stonebranch’s Terms of Use.

Version Information

Template Name	Extension Name	Version	Status

Template Name	Extension Name	Version	Status
Google Kubernetes Engine Jobs	ue-gke-jobs	2 (Current 2.1.0)	Fixes and new Features are introduced.

Refer to Changelog for version history information.

Overview

Google Kubernetes Engine (GKE) is a service for deploying, managing, and scaling Kubernetes containers on Google Cloud. Among other capabilities, GKE supports Jobs and CronJobs for task execution. Jobs are used for one-time tasks that run to completion, such as batch processing, while CronJobs enable scheduled, recurring task execution. To deploy a Job or CronJob, a Kubernetes resource definition is used. This file is typically defined inside of a YAML or JSON file and specifies the task details, container image, and execution configuration. Once applied to a GKE cluster, the Job runs to completion and terminates automatically, while CronJobs create new Job instances according to their defined schedule.

This integration provides the capability to define and deploy GKE Jobs and CronJobs on a specified cluster or delete them from a namespace.

Key Features

Feature	Description

Feature	Description
Deploy Resources	Deploy Job and CronJob resources, either declaratively or imperatively, from local YAML or JSON files, remote URLs or UAC scripts.
Create Jobs from Cronjobs	Create one-off Job resources based on existing Cronjob definitions in the cluster.
Delete Resources	Delete Job and CronJob resources via the delete action or manually through a Dynamic Command. Job resources can also be set to be deleted automatically after completion.
Pod Information Streaming	Live stream real-time updates about Kubernetes pods as they are created, modified, or terminated.
Automatic or Manual Container Information Retrieval	Option to automatically or manually retrieve information, including logs, for all containers pertaining to a specific job.

Requirements

This integration requires a Universal Agent and a Python runtime to execute the Universal Task.

Area	Details

Area	Details
Python Version	Requires Python 3.11, tested with Python 3.11.8.
Universal Agent Compatibility	Compatible with Universal Agent for Windows x64 and version >= 7.6.0.0. Compatible with Universal Agent for Linux and version >= 7.6.0.0.
Universal Controller Compatibility	Universal Controller Version >= 7.6.0.0.
Network and Connectivity	Network Connectivity towards Google Cloud Authentication Endpoints and Google Kubernetes Engine is required.
Google Kubernetes Engine	Tested with GKE 1.30.9-gke.1009000. It should be compatible with later versions as long as backwards compatibility is guaranteed on the API level.

Supported Actions

There are six Top-Level actions controlled by the Action Field:

Apply a Job definition directly to the cluster
Apply a CronJob to schedule recurring tasks
Create a new Job using a provided specification
Create a one-time Job from an existing CronJob definition
Create a new CronJob for scheduled task execution
Delete a Job or CronJob

Additional options and post-action capabilities vary depending on the chosen Action.

Action Output

Output Type	Description	Examples

Output Type

Description

Examples

EXTENSION

The EXTENSION Output provides the following information:

"exit_code", "status_description": General info regarding the Task execution. For more information, refer to the exit code table.
"invocation.fields": The task configuration used for this Task execution.
"result.resource_specification": Specification of deployed Job or CronJob, along with the name and namespace it has been deployed on. Applicable for all Actions except "Delete Resource".
"result.containers": Information relating to containers, as well as container logs pertaining to execution of deployed Job. Logs are returned for all generated pods, even if they have been restarted. Applicable if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“.
"result.errors": List of errors that might have occurred during execution.

Successful Execution of "Create Job" Action, with all Extension Output options and Wait for Success or Failure is selected

{
	"exit_code": 0,
	"status_description": "Job created successfully.",
	"invocation": {
		"extension": "ue-gke-jobs",
		"version": "2.1.0",
		"fields": { ... }
	},
	"result": {
		"containers": [
			{
				"pod_name": "pi-json-name-wxx5r",
				"container_name": "pi",
				"container_log": "3.141592653589793",
				"exit_code": "0",
				"restart_count": "0"
			},
			{
				"pod_name": "pi-json-name-wxx5r",
				"container_name": "hello-sbaro",
				"container_log": "Hello SB Team\n",
				"exit_code": "0",
				"restart_count": "0"
			}
		],
		"resource_specification": {
			"name": "pi-json-name",
			"namespace": "default",
			"active_deadline_seconds": 1800,
			"backoff_limit": 6,
			"completions": 1,
			"parallelism": 1,
			"pod_failure_policy": null,
			"pod_replacement_policy": "TerminatingOrFailed",
			"success_policy": null,
			"ttl_seconds_after_finished": 3600,
			"suspend": false
		}
	}
}

Failed Execution

{
	"exit_code": 21,
	"status_description": "Resource Naming Error: A job with the name pi-	json-name already exists in the namespace default",
	"invocation": {
	"extension": "ue-gke-jobs",
	"version": "2.1.0",
	"fields": { ... }
	},
	"result": {
		"errors": [
		"Resource Naming Error: A job with the name pi-json-name already exists in the namespace default"
		]
	}
}

STDOUT

The STDOUT output produced during the execution of the task instance.

Output can be controlled via the STDOUT Options input field. Available choices are

“Enable Pod Monitoring”, which streams information about the state of pods relating to the deployed job.
“Include Container Logs” which outputs logs for all containers tied to the deployed job.

Both options are only applicable if Wait for Success or Failure has been selected.

Truncated STDOUT, generated by the above successful task, which has all STDOUT options selected

--- Pod Monitoring Events ---
Event: ADDED
Event Time: 2025-06-06 13:13:47 +0300
Pod Name: pi-json-name-wxx5r
Phase: Pending
Conditions: Not available yet
Start Time: None

Event: MODIFIED
Event Time: 2025-06-06 13:13:47 +0300
Pod Name: pi-json-name-wxx5r
Phase: Pending
Conditions: PodScheduled
Start Time: None

Event: MODIFIED
Event Time: 2025-06-06 13:13:47 +0300
Pod Name: pi-json-name-wxx5r
Phase: Pending
Conditions: PodReadyToStartContainers, Initialized, Ready, ContainersReady, PodScheduled
Start Time: 2025-06-06 10:13:47+00:00
	Container: hello-sbaro
		Waiting:
			Reason: PodInitializing

	Container: pi
		Waiting:
			Reason: PodInitializing

Event: MODIFIED
Event Time: 2025-06-06 13:13:48 +0300
Pod Name: pi-json-name-wxx5r
Phase: Pending
Conditions: PodReadyToStartContainers, Initialized, Ready, ContainersReady, PodScheduled
Start Time: 2025-06-06 10:13:47+00:00
	Container: hello-sbaro
		Waiting:
			Reason: PodInitializing

	Container: pi
		Waiting:
			Reason: PodInitializing

[...]

Pod pi-json-name-wxx5r has reached phase 'Succeeded' but is pending deletion due to finalizers
Event: MODIFIED
Event Time: 2025-06-06 13:13:51 +0300
Pod Name: pi-json-name-wxx5r
Phase: Succeeded
Conditions: PodReadyToStartContainers, Initialized, Ready, ContainersReady, PodScheduled
Start Time: 2025-06-06 10:13:47+00:00
	Container: hello-sbaro
		Terminated:
			Exit Code: 0
			Reason: Completed
			Started at: 2025-06-06 10:13:48+00:00
			Finished at: 2025-06-06 10:13:48+00:00

	Container: pi
		Terminated:
			Exit Code: 0
			Reason: Completed
			Started at: 2025-06-06 10:13:48+00:00
			Finished at: 2025-06-06 10:13:48+00:00

Pod pi-json-name-wxx5r has reached phase 'Succeeded'

Pod pi-json-name-wxx5r successfully completed, 1/1 pods completed.

Job pi-json-name execution finished, status is 'Complete'.

--- Logs for containers generated for job pi-json-name at 2025-06-06 13:13:52 +0300 ---
Logs for pod pi-json-name-wxx5r
	Logs for container pi:
		3.141592653589793
	Logs for container hello-sbaro:
		Hello SB Team

Configuration Examples

Example: Apply a Job stored as a UAC Script

Apply a Job to a specific GKE cluster, authenticating with a Service Account Key File, with the Resource Definition provided as a UAC script. The EXTENSION Output includes detailed information about the Job specification.

Example: Apply a CronJob stored as a Local File

Apply a CronJob to a specific GKE cluster, authenticating with a Service Account Key File, with the Resource Definition provided as a Local File. Since the namespace is not specified in the resource definition file, the Namespace field input value will be used to determine where the CronJob will be applied.

Example: Create a Job defined on an external HTTP Link and wait for Job completion

Deploy a Job to a specific GKE cluster, authenticating with a Service Account Key File, with the Resource Definition provided as a HTTP Link. The Job's execution is monitored until completion for up to 1500 seconds, with any information related to the pods tied to the Job printed on STDOUT. If the Job executes successfully, it will be deleted from the designated namespace. Once the Job completes, the container logs and are printed on STDOUT. The EXTENSION Output includes detailed information about the Job specification and the containers created to complete it.

Example: Create a Job from an existing CronJob and wait for Job completion

Deploy a Job to a specific GKE cluster, authenticating with a Service Account Key File, using the selected CronJob. The Job's execution is monitored until completion for up to 500 seconds, retrieving the Container Logs and Information if the Job fails. Once the Job completes, it will be deleted from the designated namespace.

Example: Create a CronJob stored as a UAC script using Impersonation

Deploy a CronJob to a specific GKE cluster, authenticating with a Service Account Key File using Impersonation, with the Resource Definition provided as a UAC Script. The EXTENSION Output includes detailed information about the CronJob specification.

Example: Delete Resource

Delete Resource of Resource Kind “Job“ on the GKE cluster, retrieving the Job name using the Dynamic Choice field from specific namespace.

Input Fields

Name	Type	Description	Version Information

Name	Type	Description	Version Information
Action	Choice	The action performed upon Task execution. The following options are available. Apply Job (default) Apply CronJob Create Job Create Job from CronJob Create CronJob Delete Resource	Introduced in 1.0.0
Authentication Method	Choice	The way the Google Cloud account will be verified. The following option is available. Service Account Key	Introduced in 1.0.0
Service Account Key	Credential	The UAC Credential containing the Google Cloud Service Account Key to be used for authentication. The contents of the Service Account Key should be input on the Token field of a Resolvable UAC Credential. Only available if Authentication Method is set to "Service Account Key".	Introduced in 1.0.0
Impersonate Service Account	Checkbox	Specifies whether a different Service Account should be impersonated.	Introduced in 1.0.0
Target Service Account	Text	The Service Account to impersonate. IAM roles required for creating short-lived tokens must be configured beforehand. Only available if Impersonate Service Account is selected.	Introduced in 1.0.0
Cluster	Dynamic Choice	The Cluster to deploy or delete the resource on.	Introduced in 1.0.0
Namespace	Text	The namespace in which resources are deployed or deleted. In the case of resource deployment, this field is used only as a fallback in case a ‘namespace’ value is not specified inside of the Resource Definition.	Introduced in 1.0.0
Resource Definition Source	Choice	Specifies how to retrieve the definition of the resource to be deployed. The content of the file should be in YAML or JSON format and should contain only one resource definition. The following options are available. UAC Script Local File HTTP Link Only available if Action is set to “Apply Job”, “Apply CronJob“, “Create Job“, or “Create CronJob“.	Introduced in 1.0.0
Resource Definition Script	Script	UAC Script containing the definition of the resource to be deployed. Only available if Resource Definition Source is set to "UAC Script".	Introduced in 1.0.0
Resource Definition File Path	Text	Local path to the resource definition file. Only available if Resource Definition Source is set to "Local File".	Introduced in 1.0.0
Resource Definition Link	Text	HTTP link to file to use as resource definition. If authentication is required to retrieve the file, it should be part of the link provided. Environment variables in the URL are resolved; The `GIT_TOKEN` environmental variable, if set, is used as a Bearer token for authorization. Only available if Resource Definition Source is set to “HTTP Link”.	Introduced in 1.0.0
CronJob to Use	Dynamic Choice	The CronJob to use when creating a Job from CronJob. Available and required if Action is set to “Create Job from CronJob“.	Introduced in 2.0.0
New Job Name	Text	The name that the created Job should have. Available and required if Action is set to “Create Job from CronJob“	Introduced in 2.0.0
Resource Kind	Choice	The kind of resource to be deleted. The following options are available. Job (default) CronJob Available and required if Action is set to “Delete Resource“.	Introduced in 2.0.0
Resource Name	Dynamic Choice	The name of the resource to be deleted. Only available if Action is set to "Delete Resource".	Introduced in 1.0.0
Wait for Success or Failure	Checkbox	If selected, makes the Task Instance wait until the job had been completed. Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“. Note: When applying to an existing Job resource, monitoring functionality is disabled by default.	Introduced in 1.0.0
Delete Job Options	Choice	Specifies the conditions under which the deployed Job should be deleted. The following options are available. Do not Delete Job Delete Job on Successful Execution (default) Delete Job Regardless of Execution Status Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“ and Wait for Success or Failure is selected.	Introduced in 1.0.0
Timeout Seconds	Integer	The maximum number of seconds to monitor for Job completion. Leave empty or set to 0 to monitor indefinitely. Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“ and Wait for Success or Failure is selected.	Introduced in 1.0.0
Retrieve Container Information Only on Failure	Checkbox	Specifies whether container information is only retrieved when the Job fails. This option does nothing if "Include Container Logs" or "Include Container Information" options are not requested as part of the Standard Output Options or Extension Output Options, respectively. Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“ and Wait for Success or Failure is selected.	Introduced in 1.0.0
Extension Output Options	Multiple Choice	Controls whether additional information is printed to Extension Output. The following options are available. Include Resource Specification and Metadata Include Container Information (Only applicable if Wait for Success or Failure is selected) Available for all Actions except“Delete Resource“.	Introduced in 1.0.0
STDOUT Options	Multiple Choice	Controls whether additional information is printed to Standard Output. The following options are available. Include Container Logs Enable Pod Monitoring Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“ and Wait for Success or Failure is selected.	Introduced in 1.0.0

Output Fields

The following output-only fields provide better visibility during the execution of Universal Task Instances.

Name	Type	Description	Version Information

Name	Type	Description	Version Information
Resource Name	Text	Name of deployed resource. Applicable for all Actions except“Delete Resource“.	Introduced in 1.0.0
Resource Namespace	Text	Namespace the resource has been deployed on. Applicable for all Actions except“Delete Resource“.	Introduced in 1.0.0
Job Status	Text	The latest status of the Job. Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“ and Wait for Success or Failure is selected.	Introduced in 2.0.0
Last Updated On	Text	The timestamp when the Task Instance was last updated. Available if Action is set to “Apply Job”, “Create Job“, or “Create Job from CronJob“ and Wait for Success or Failure is selected.	Introduced in 2.0.0

Environment Variables

Environment Variables can be set from the Environment Variables Task definition table. They can be useful for authentication purposes when retrieving the Resource Definition via "HTTP Link".

Environment Variable Name	Description	Version Information

Environment Variable Name	Description	Version Information
GIT_TOKEN	If specified, this environment variable will be used as a Bearer token in the Authorization header when making the request to retrieve a Resource Definition via HTTP Link.	Introduced in 1.0.0
UE_HTTP_TIMEOUT	This environment variable defines the global timeout value for all requests. It affects both connection and read timeouts, preventing requests from hanging. Accepted values: Positive integer (in seconds). If the environment variable is not set, or is set to an invalid value, the timeout will default to 60 seconds.	Introduced in 2.0.0
UE_RESOURCE_DEFINITION_LINK_GIT_PROVIDER	This environment variable is used only when `GIT_TOKEN` is passed and defines which Git provider to use when fetching Kubernetes job resource definitions. It ensures the correct Authorization header is generated for the selected provider. Supported values are: `github` `gitlab` `bitbucket` `azure_devops` If the environment variable is not set, `github` is used by default.	Introduced in 2.1.0
UE_JOB_FINALIZATION_POLL_INTERVAL	This environment variable defines the interval (in seconds) between successive checks for job finalization after the pods have completed. It controls the polling frequency to ensure that job completion is detected reliably without overwhelming the server or agent resources. Accepted values: Positive integer (in seconds). If the environment variable is not set, or is set to an invalid value, the polling interval will default to 1 second.	Introduced in 2.1.0

Cancelation and Rerun

In case of cancellation, the extension will stop monitor the job’s current status. If the deployment process has already been initiated, it will not be impacted in any way. Additionally, in case Extension Output Options have been specified, information that has been produced at the point of cancellation will be returned.

In the case of a Task Rerun, the extension follows the same execution flow as the initial execution. As such, resources will be deployed or deleted, depending on the specified input fields, with no regard to the current condition of the GKE cluster.

Dynamic Commands

Users can Interact with Task Instances through Dynamic Commands. The following Dynamic Commands are supported.

Name	Description	Version Information

Name