Also- the env field you mention applies to celery workers, which monitor user code but don’t actually execute it. The celery-k8s launcher also takes envConfigMaps, which will apply to the k8s jobs that actually run user code.
Assuming the env is for your code (as opposed to Celery config), you’ll want to use the envConfigMaps field
One way to manage these configmaps is with our helm chart, under
04/06/2021, 7:40 AM
Hi @johann, thanks for your quick response. So I need to pipe the values through the kubernetes environment (configmaps/secrets) for them to reach the k8sRunLauncher-created pods. I originally thought that the k8sRunLauncher could be used as the userDeployment where there is also an