API reference

JobManager

class kubernetes_job.JobManager(k8s_client: kubernetes.client.api_client.ApiClient, k8s_job_spec: [<class 'dict'>, <class 'str'>], namespace: str = 'default')

Kubernetes JobManager

Parameters

• k8s_client – Kubernetes OpenAPI client

• k8s_job_spec – dict or path to YAML file containing the spec for the job worker

• namespace – Kubernetes k8s_namespace (default: ‘default’)

clean_jobs(field_selector=None, label_selector=None)

Clean up completed jobs

Parameters

• field_selector – A selector to restrict the list of returned objects by their fields. Defaults to everything.

• label_selector – A selector to restrict the list of returned objects by their labels. Defaults to everything.

create_job(func, *func_args, **func_kwargs) → kubernetes.client.models.v1_job.V1Job

Create a job

Parameters

• func – Function pointer

• func_args – Args to submit to the function

• func_kwargs – Kwargs to submit to the function

Returns

V1Job

delete_job(job: [<class 'str'>, <class 'kubernetes.client.models.v1_job.V1Job'>], grace_period_seconds: int = 0, propagation_policy: str = 'Background') → kubernetes.client.models.v1_status.V1Status

Delete a Job

Parameters

• job – Name or V1Job instance

• grace_period_seconds – (default: 0)

• propagation_policy – (default: ‘Background’)

Returns

V1Status

static execute_job(job_func_def: Optional[str] = None)

Execute the JobFuncDef specified in the func_spec

Parameters

job_func_def – Serialized job definition

Returns

Job function return value (if any)

list_jobs(field_selector=None, label_selector=None) → Iterator[kubernetes.client.models.v1_job.V1Job]

List job objects

Parameters

• field_selector – A selector to restrict the list of returned objects by their fields. Defaults to everything.

• label_selector – A selector to restrict the list of returned objects by their labels. Defaults to everything.

Returns

Iterator of V1Job

read_job(job: [<class 'str'>, <class 'kubernetes.client.models.v1_job.V1Job'>]) → kubernetes.client.models.v1_job.V1Job

Read the status of the specified Job

Parameters

job – Name or V1Job instance

Returns

V1Job

Helpers

kubernetes_job.job_name(job: [<class 'str'>, <class 'kubernetes.client.models.v1_job.V1Job'>]) → str

Return the name of a job

kubernetes_job.job_status(job: kubernetes.client.models.v1_job.V1Job) → str

Return SUCCEEDED, FAILED, ACTIVE, or PENDING, depending on the status of the job

kubernetes_job.is_completed(job: kubernetes.client.models.v1_job.V1Job)

Return True if the job has completed (either failed or succeeded)

kubernetes_job.is_succeeded(job: kubernetes.client.models.v1_job.V1Job)

Return True if the job has succeeded

kubernetes_job.is_failed(job: kubernetes.client.models.v1_job.V1Job)

Return True if the job is failed

kubernetes_job.is_active(job: kubernetes.client.models.v1_job.V1Job)

Return True if the job is active (running)

kubernetes_job.current_job =kubernetes_job.job_func_def.JobFuncDef

Current JobFuncDef when executing a Kubernetes-job (as runner), otherwise None.

job_func_def

job_func_def contains helper classes for the serialization and execution of the function call.

class kubernetes_job.job_func_def.JobFuncDef(func, args=None, kwargs=None, meta: Optional[kubernetes_job.job_func_def.JobMeta] = None)

Helper class to hold the job function definition

Parameters

• func – Pointer to the job function

• args – Args for the job function

• kwargs – Kwargs for the job function

• meta – Metadata for the job

args = None

Args for the job function

dump() → str

Dump the job function definition to a base64 string

execute()

Execute the job function

func = None

Pointer to the job function

kwargs = None

Kwargs for the job function

static load(s: str) → kubernetes_job.job_func_def.JobFuncDef

Load the job function definition from a base64 string

meta: kubernetes_job.job_func_def.JobMeta = None

Metadata for the job

class kubernetes_job.job_func_def.JobMeta

Helper class to hold job meta information

dt_scheduled: <module ‘datetime’ from ‘/home/docs/.pyenv/versions/3.7.9/lib/python3.7/datetime.py’> = datetime.datetime(1, 1, 1, 0, 0)

Job scheduled datetime

host: str = '[HOST]'

Host responsible for spawning the job

name: str = '[JOB-NAME]'

Unique job name