Failure to remove resources used by completed jobs/pods can lead to resource exhaustion in the Kubernetes cluster. This commit adds support for setting a ttlSecondsAfterFinished parameter in the Kubernetes Job sepc. This parameter specifies the amount of time (in seconds) for which the resources used by completed job/pod will persist after it which it will be garbage collected by the Kube control manager.
The newly added parameter is optional and only set if it passed in through the config argument to the make_job_spec method.
Failure to remove resources used by completed jobs/pods can lead to resource exhaustion in the Kubernetes cluster. This commit adds support for setting a ttlSecondsAfterFinished parameter in the Kubernetes Job sepc. This parameter specifies the amount of time (in seconds) for which the resources used by completed job/pod will persist after it which it will be garbage collected by the Kube control manager.
The newly added parameter is optional and only set if it passed in through the
config
argument to themake_job_spec
method.Purpose of PR
Parts of the app this will impact
Todos
Additional information
Related links